# Algenta

## Algenta

- [What Algenta is](https://docs.algenta.ai/overview.md): Algenta is compiler-driven, self-hosted AI infrastructure — 6,000+ deterministic functions on custom Mojo kernels, reached over the HTTP API, the Python and TypeScript SDKs, or MCP. Start here.
- [Install](https://docs.algenta.ai/getting-started/install.md): Install the published Algenta clients or run the self-hosted engine. Pick your path, install in one command where a public package exists, and verify it.
- [Quickstart](https://docs.algenta.ai/getting-started/quickstart.md): Run Algenta end to end in minutes — install the integrated service, mint an API key, and make your first decision call.
- [Try it without a key](https://docs.algenta.ai/getting-started/playground.md): Try the Algenta engine with a single POST /v1/playground call — no sign-up and no API key. Run a decision, optimization, simulation, or forecast in one request.
- [Core concepts](https://docs.algenta.ai/getting-started/concepts.md): The core Algenta model — one integrated engine, governed data, deterministic decisions, capability routing, and explicit deployment boundaries.
- [Authentication & API keys](https://docs.algenta.ai/getting-started/authentication.md): Create and use Algenta API keys. Mint a key, send it as a Bearer token, and confirm an authenticated request succeeds.
- [Verify the bundled engine](https://docs.algenta.ai/engine-runtime/bundled-runtime.md): Verify the bundled Algenta compute runtime without installing, starting, or managing a separate engine process.
- [Engine health & verification](https://docs.algenta.ai/engine-runtime/health.md): Confirm that the API, dependencies, and bundled compute runtime are healthy through supported HTTP probes.
- [Engine catalog](https://docs.algenta.ai/engine-runtime/catalog.md): The Algenta engine catalog — every engine, its slug, API path, tier, and use cases, plus the three endpoints that enumerate them at runtime.
- [Runtime libraries](https://docs.algenta.ai/engine-runtime/runtime-libraries.md): The runtime library surface — list, health-check, and execute individual runtime-backed functions over HTTP, the de CLI, or MCP. Off by default; an operator enables the HTTP routes.
- [Runtime library catalog](https://docs.algenta.ai/engine-runtime/runtime-library-catalog.md): The full catalog of Algenta runtime-library modules — vision, LLM, ML, tensors, statistics, signal, crypto, and more — callable through /v1/libraries.
- [Kernel inventory](https://docs.algenta.ai/engine-runtime/kernel-inventory.md): The public inventory of the engine's CUDA and Metal kernels — what ships, per backend, per model family, and the evidence behind each module.
- [Declared capabilities](https://docs.algenta.ai/engine-runtime/capabilities.md): Ask the engine what each kernel actually does — deterministic or not, seeded or not, stateful or not, what it costs — and be told plainly when a kernel has not said.
- [Batched and chained execution](https://docs.algenta.ai/engine-runtime/coarse-execution.md): Run many kernels — or a whole chain — in one call instead of one call each.
- [Reproducibility](https://docs.algenta.ai/engine-runtime/reproducibility.md): Getting the same answer twice — chunk-invariant reductions, one canonical spelling for every float, and a replay manifest that says whether two runs were really the same run.
- [Metal LLM decode](https://docs.algenta.ai/gpu-and-llm-inference/qwen3-metal-decode.md): Greedy decode on Apple Silicon through the int8 Metal GPU kernel path — call surface, checkpoint preparation, refusals, verification. Currently tested on Qwen3 models.
- [Model support matrix](https://docs.algenta.ai/gpu-and-llm-inference/model-support-matrix.md): Model support as declared fact ranges — the capability table, the conform CLI, conformance-proven models, and the typed-refusal boundary.
- [GPU support](https://docs.algenta.ai/gpu-and-llm-inference/gpu-support.md): GPU backend support matrix — Metal (default serve path), NVIDIA CUDA (hardware-verified kernel layer and LLM session), ROCm status, and per-lane requirements.
- [CUDA backend](https://docs.algenta.ai/gpu-and-llm-inference/cuda-backend.md): Run the CUDA backend — availability, NVIDIA driver and architecture requirements, the callable entry points, determinism gates, verified model shapes, and the TypeScript SDK path.
- [Data integrity and encoding](https://docs.algenta.ai/compute-and-kernel-primitives/data-integrity-and-encoding.md): Checksums, encoding, keyed hashing, key derivation, a stream cipher and a Merkle tree — compiled primitives, not a general-purpose library wrapper.
- [Config and data-parsing kernels](https://docs.algenta.ai/compute-and-kernel-primitives/config-and-data-parsing-kernels.md): TOML and YAML config parsing with dot-notation lookups, an RFC 4180 CSV reader, a columnar DataFrame factory, and the zero-Python JSON codec that the engine's own dispatch layer decodes every runtime.
- [Compression kernels](https://docs.algenta.ai/compute-and-kernel-primitives/compression-kernels.md): Delta, Huffman, LZ77 and run-length encoding as compiled kernels — four classic lossless algorithms, each matched to one specific shape of redundancy.
- [Tensor buffer primitives](https://docs.algenta.ai/compute-and-kernel-primitives/tensorx-kernels.md): Dtype-tagged byte buffers, elementwise ops, wide-accumulating reductions, N-D shape/broadcasting and low-bit quantization — the flat tensor substrate other kernel families decode through.
- [Normalization kernels](https://docs.algenta.ai/compute-and-kernel-primitives/normalization-kernels.md): LayerNorm and RMSNorm forward/backward kernels over flat row-major buffers, plus the shared contract that validates their row shapes and epsilon.
- [Attention kernels](https://docs.algenta.ai/compute-and-kernel-primitives/attention-kernels.md): Scaled dot-product attention, flash attention, causal/sliding-window masking, RoPE, and analytic backward gradients, in flat row-major layouts with grouped-query support.
- [MLP block kernels](https://docs.algenta.ai/compute-and-kernel-primitives/mlp-block-kernels.md): Gated feed-forward activations (SwiGLU and both GEGLU conventions), token embedding lookup, and the shared feed-forward/vocabulary shape contract, forward and backward, in flat row-major layouts.
- [GEMM kernels](https://docs.algenta.ai/compute-and-kernel-primitives/gemm-kernels.md): Cache-blocked SIMD dense GEMM/GEMV in Float64, NF4/int4/int8 quantized matmul with dequantization fused into the reduction loop, and the shared status/dimension/work-budget contract behind both.
- [GPU acceleration kernels](https://docs.algenta.ai/compute-and-kernel-primitives/gpu-acceleration-kernels.md): Backend selection contract, host device capability probing without compiling a shader, and a threadgroup-tiled GEMM kernel on the Metal GPU.
- [LLM inference kernels](https://docs.algenta.ai/compute-and-kernel-primitives/llm-inference-kernels.md): A parallel decode-shaped matrix-vector product, KV-cache growth, greedy argmax, whole-buffer residual add, and the shared status vocabulary that a real autoregressive decode loop composes them into.
- [Attention variant kernels](https://docs.algenta.ai/compute-and-kernel-primitives/transformer-attention-variant-kernels.md): Scalar reference formulas for scaled dot-product attention, ten named sparse/long-context attention techniques, flash attention's online-softmax building blocks, and five families of positional encodi
- [Transformer blocks and normalization](https://docs.algenta.ai/compute-and-kernel-primitives/transformer-blocks-and-normalization-reference-kernels.md): Scalar reference formulas for assembling a transformer block (residual, pre/post-norm, causal mask, FFN), a broad normalization catalog beyond LayerNorm/RMSNorm, general non-transformer NN primitives,
- [KV-cache and generation kernels](https://docs.algenta.ai/compute-and-kernel-primitives/kv-cache-and-autoregressive-generation-kernels.md): Scalar formulas for KV-cache memory and eviction (plain and paged), sampling and generation-loop decisions, speculative-decoding accept-rate math, and continuous-batching scheduling -- the arithmetic
- [Multimodal and inference operations](https://docs.algenta.ai/compute-and-kernel-primitives/multimodal-and-inference-operations-kernels.md): Scalar analytic formulas for multimodal fusion primitives, inference-serving capacity and latency estimates, per-request cost accounting, and prompt-injection/output-safety scoring.
- [Loss functions and resident-context kernels](https://docs.algenta.ai/compute-and-kernel-primitives/loss-functions-and-resident-context-kernels.md): Chunked cross-entropy that never materializes the full logit tensor, and the host-side capacity and dispatch-budget arithmetic for a resident GPU decode session, answered with no device required.
- [Neural network primitives and data-pipeline kernels](https://docs.algenta.ai/compute-and-kernel-primitives/neural-network-primitives-and-data-pipeline-kernels.md): Eleven unrelated small libraries closing out the kernel-documentation backlog -- scalar activation functions, cellular automata, a string-similarity column-name ranker, two distinct data-pipeline util
- [Core linear algebra](https://docs.algenta.ai/numerical-and-mathematical-computing/core-linear-algebra-kernels.md): Dense matrix arithmetic, LU/QR/Cholesky decompositions, 2×2 eigenvalues and power iteration, and elementwise matrix operations, split across two independent in-tree matrix representations.
- [Exact numerics and graph kernels](https://docs.algenta.ai/numerical-and-mathematical-computing/exact-numerics-and-graph-kernels.md): Double-double error-free-transform arithmetic and compensated Householder least squares, plus explicit-graph traversal, shortest paths, MST, and union-find.
- [Sparse structures & tensor primitives](https://docs.algenta.ai/numerical-and-mathematical-computing/sparse-and-tensor-kernels.md): Coordinate-format sparse matrices and paired-array sparse vectors, a dense 1D tensor type, and the standalone scalar index/accumulation formulas an N-D tensor implementation composes.
- [Probability distributions & dependence](https://docs.algenta.ai/numerical-and-mathematical-computing/probability-distributions-and-dependence-kernels.md): Sampling and MLE/method-of-moments fitting for twenty-six probability distributions, core probability theory (Bayes, combinatorics, entropy), a Gaussian copula with rank-correlation measures, and Chol
- [Statistical computing](https://docs.algenta.ai/numerical-and-mathematical-computing/statistical-computing-kernels.md): Descriptive statistics with a bundled introsort, bootstrap and jackknife resampling for confidence intervals and bias, and Monte Carlo simulation of random walks, diffusions, and point processes.
- [Calculus & differential equations](https://docs.algenta.ai/numerical-and-mathematical-computing/calculus-and-differential-equation-kernels.md): Interpolation between data points, quadrature over sampled data, initial-value ODE solvers (Euler, RK4), and explicit/Jacobi finite-difference PDE solvers with their CFL stability limits.
- [Polynomial, root-finding & number theory](https://docs.algenta.ai/numerical-and-mathematical-computing/polynomial-root-finding-and-number-theory-kernels.md): Dense-coefficient polynomial arithmetic and calculus, five classical root-finding algorithms run against a fixed menu of test functions, and integer number theory -- primality, GCD, modular exponentia
- [Combinatorics, convex optimization & scipy-compat](https://docs.algenta.ai/numerical-and-mathematical-computing/combinatorics-convex-optimization-and-scipy-compat-kernels.md): Combinatorial search heuristics, an unconstrained convex-optimization sandbox, a frozen two-variable LP solver, quasi-random sampling, and the scalar formulas behind scipy.optimize and scipy.stats.
- [Signal transforms & processing](https://docs.algenta.ai/numerical-and-mathematical-computing/signal-transforms-and-processing-kernels.md): Radix-2 FFT and Haar wavelet transforms, two independently-implemented frequency-domain feature modules, window/peak/zero-crossing signal utilities, whole-column statistical data transforms, and the f
- [Dimensionality, information theory & complex systems](https://docs.algenta.ai/numerical-and-mathematical-computing/dimensionality-information-theory-and-complex-systems-kernels.md): Six unrelated math primitives grouped by name-similarity rather than shared subject -- business unit-consistency checking, PCA, Shannon information theory, special functions, chaotic dynamical systems
- [Quantum computing & formal logic](https://docs.algenta.ai/numerical-and-mathematical-computing/quantum-computing-and-formal-logic-kernels.md): Single-qubit state-vector simulation plus closed-form quantum-computing estimators, propositional/boolean logic and truth tables, fuzzy-set membership and defuzzification, gate-level digital circuit s
- [Geometry and simulation-physics kernels](https://docs.algenta.ai/numerical-and-mathematical-computing/geometry-and-simulation-physics-kernels.md): 2D Euclidean geometry, rigid-body game-engine kinematics, chemical reaction-rate kinetics, and point-cloud topological data analysis -- four self-contained modules that share no code with each other o
- [Time-series model selection](https://docs.algenta.ai/machine-learning/ml-model-selection.md): Walk-forward, purged, k-fold and temporal splits for leakage-free time-series model selection.
- [Statistical model comparison](https://docs.algenta.ai/machine-learning/ml-model-comparison.md): Paired bootstrap CIs, exact McNemar and Diebold-Mariano tests that decide whether a candidate model is genuinely better than the champion.
- [Probability calibration](https://docs.algenta.ai/machine-learning/ml-calibration.md): Isotonic (PAV) and z-score affine calibration with reliability diagnostics for turning raw model scores into trustworthy probabilities.
- [Bayesian methods, calibration, and clustering metrics](https://docs.algenta.ai/machine-learning/bayesian-methods-and-calibration-kernels.md): Conjugate Bayesian updates and Monte Carlo A/B testing, general-purpose variational/Laplace inference primitives, probability calibration, and cluster-quality metrics across four independent kernel mo
- [Dense 2D matrices (Matrix2D)](https://docs.algenta.ai/machine-learning/ml-tensor2d.md): Dense 2-D matrix type with matmul, matvec and outer products backing the trainable-model kernels.
- [LSTM and GRU cells](https://docs.algenta.ai/machine-learning/ml-sequence-cells.md): LSTM and GRU cell forward passes with exact gate math for sequence models.
- [LSTM sequence training (BPTT)](https://docs.algenta.ai/machine-learning/ml-sequence-training.md): Full BPTT training loop for LSTM regressors, gradient-checked against finite differences.
- [Bit-exact weight serialization](https://docs.algenta.ai/machine-learning/ml-serialization.md): Bit-exact float64 weight serialization for saving and restoring trained kernels.
- [Gradient-boosted regression trees](https://docs.algenta.ai/machine-learning/ml-gradient-boosting.md): Gradient-boosted regression trees with exact SSE-gain splits, subsampling and shrinkage.
- [Gaussian mixture models](https://docs.algenta.ai/machine-learning/ml-mixture-models.md): Gaussian mixture models fit by log-space EM with BIC/AIC model selection.
- [Hidden Markov models](https://docs.algenta.ai/machine-learning/ml-hmm.md): Hidden Markov models — forward/backward likelihood, Viterbi decoding and Baum-Welch training.
- [Imbalanced-data resampling](https://docs.algenta.ai/machine-learning/ml-imbalance.md): SMOTE, random over/under-sampling and balanced class weights for skewed datasets.
- [Model ensembling](https://docs.algenta.ai/machine-learning/ml-ensembles.md): Stacking and weighted voting ensembles over arbitrary base-model predictions.
- [Conformal prediction intervals](https://docs.algenta.ai/machine-learning/ml-conformal.md): Split-conformal prediction intervals with finite-sample coverage guarantees.
- [Model explainability](https://docs.algenta.ai/machine-learning/ml-explainability.md): KernelSHAP and permutation importance for explaining any model's predictions.
- [Gaussian process regression](https://docs.algenta.ai/machine-learning/ml-gaussian-processes.md): Gaussian process regression with RBF and Matern kernels, exact posterior mean and variance.
- [Hyperparameter search schedules](https://docs.algenta.ai/machine-learning/ml-automl.md): Hyperband successive-halving schedules and Latin-hypercube search plans for hyperparameter optimization.
- [Fit/transform pipelines](https://docs.algenta.ai/machine-learning/ml-pipelines.md): Composable fit/transform preprocessing pipelines (standardize, min-max, chained) with exact inverses.
- [Trainable linear models](https://docs.algenta.ai/machine-learning/ml-linear-models.md): Trainable logistic regression and elastic-net linear models with exact coordinate-descent math.
- [Symmetric eigendecomposition and thin SVD](https://docs.algenta.ai/machine-learning/ml-eigendecomposition.md): Jacobi symmetric eigendecomposition and thin SVD for full-rank spectral computation.
- [AR/MA estimation (ARIMA building blocks)](https://docs.algenta.ai/machine-learning/ml-arima.md): Durbin-Levinson AR estimation, innovations MA fitting and ARIMA forecasting building blocks.
- [Feature hashing (the hashing trick)](https://docs.algenta.ai/machine-learning/ml-feature-hashing.md): Seeded FNV-1a feature hashing for fixed-width vectorization of unbounded feature names.
- [Partial least squares regression (PLS1)](https://docs.algenta.ai/machine-learning/ml-pls.md): NIPALS PLS1 regression for collinear and wide (p > n) feature matrices, sklearn-matched.
- [Trainable SVM classifier (simplified SMO)](https://docs.algenta.ai/machine-learning/ml-svm.md): Trainable soft-margin SVM classifier via simplified SMO with linear and RBF kernels.
- [Batched chain greeks](https://docs.algenta.ai/machine-learning/ml-chain-greeks.md): SIMD-batched Black-Scholes price and greeks across a whole option chain in one call, matching the scalar pricing functions to 1e-9.
- [Classical ML algorithms](https://docs.algenta.ai/machine-learning/ml-classical-algorithms.md): Distance functions, k-means and k-NN, Gaussian Naive Bayes, preprocessing transforms, evaluation metrics, and a bagged decision-stump random forest across seven independent kernel modules.
- [Regression and classification kernels](https://docs.algenta.ai/machine-learning/regression-and-classification-kernels.md): Closed-form OLS/ridge regression, a recursive CART regression tree, four SVM kernel functions with Gram-matrix diagnostics, and ten scalar loss functions for classical supervised learning.
- [Autograd and training stability](https://docs.algenta.ai/machine-learning/autograd-and-training-stability-kernels.md): Reverse-mode autodiff graph bookkeeping, per-operation gradient formulas, ten optimizer update rules, and the loss-scaling/NaN-guard/numerically-safe primitives that keep a training loop from divergin
- [LoRA adapters, checkpoint loading & distillation](https://docs.algenta.ai/machine-learning/lora-distillation-and-weight-loading-kernels.md): Per-element LoRA/DoRA/prefix/prompt adapter math, sharded and quantized checkpoint-loading arithmetic, and classical distillation-column design equations (not neural-network knowledge distillation).
- [Mixture-of-experts and routing](https://docs.algenta.ai/machine-learning/mixture-of-experts-and-routing-kernels.md): Sparse mixture-of-experts math (gating, capacity, load-balance and z-loss) plus the separate, higher-level layer that routes a whole request to a model, precision, or backend by cost and latency.
- [Model lifecycle and interop](https://docs.algenta.ai/machine-learning/model-lifecycle-and-interop-kernels.md): ABI version and layout arithmetic, safetensors/GGUF/ONNX/HuggingFace interop math, data/tensor/pipeline parallelism formulas, and checkpoint/quantization sizing, across four independent Float64 kernel
- [Sequence and state-space models](https://docs.algenta.ai/machine-learning/sequence-and-state-space-model-kernels.md): A 2-layer LSTM training loop with full BPTT and Adam, the per-element linear recurrence math behind Mamba/S4/RetNet-style state-space models, and the scalar heuristics a stateful agent runtime scores
- [RLHF and generative training](https://docs.algenta.ai/machine-learning/rlhf-and-generative-training-kernels.md): Preference-optimization losses for RLHF/DPO alignment, diffusion noise schedules and denoising steps, InfoNCE-style contrastive losses, one-sample-at-a-time online learning updates, and Pareto/TOPSIS
- [Feedback loops and feature engineering](https://docs.algenta.ai/machine-learning/feedback-loops-and-feature-engineering-kernels.md): Scalar scoring functions for the data flywheel (drift, retraining triggers, RLHF reward, calibration) and the scaling/encoding/selection utilities that build model inputs from raw columns.
- [Optimization & decision programs](https://docs.algenta.ai/optimization-and-simulation/optimization.md): Solve linear, integer, assignment and network-flow programs with compiled kernels on your own machine — and be told when no answer exists.
- [Optimization benchmark plane](https://docs.algenta.ai/optimization-and-simulation/optimization-model-formulation.md): A flat typed LP/MIP formulation container, its refusal-first validator and x>=0 lowering, a solve bridge that re-audits every answer in the caller's own coordinates, and a judge that scores a solved o
- [Routing, scheduling & trade-offs](https://docs.algenta.ai/optimization-and-simulation/optimization-heuristics.md): Vehicle routes, schedules, convex quadratic programs, Pareto fronts and decisions under uncertainty — searched on your own machine, reporting the bound you have rather than claiming an optimum.
- [General-purpose optimization kernels](https://docs.algenta.ai/optimization-and-simulation/optimization-suite-kernels.md): Constrained nonlinear optimization by penalty/barrier/Lagrangian methods, four metaheuristic search strategies, scalar root-finding and minimization on polynomials, and ten modern deep-learning optimi
- [Discrete-event simulation & queueing](https://docs.algenta.ai/optimization-and-simulation/discrete-event-simulation.md): Discrete-event simulation with a totally-ordered event calendar, resource pools with priority queues, warm-up detection that reports what it discarded
- [Strategy comparison & simulation](https://docs.algenta.ai/optimization-and-simulation/strategy-simulation.md): Compare strategies on common random numbers, rank them with the multiplicity corrected, and get "no significant difference" as an answer instead of a coin flip.
- [Contest simulation & service timing](https://docs.algenta.ai/optimization-and-simulation/contest-strategy.md): Simulate N competitors over discrete rounds and compare service-timing options under common random numbers, with simultaneous intervals and an explicit abstain.
- [Reproducible evolutionary search](https://docs.algenta.ai/optimization-and-simulation/reproducible-search.md): Evolutionary parameter search whose every draw is addressable, so a fitted result replays bit-exactly on any machine and any core count.
- [Reinforcement learning and stochastic simulation kernels](https://docs.algenta.ai/optimization-and-simulation/reinforcement-learning-and-stochastic-simulation-kernels.md): Tabular Q-learning, SARSA and value iteration over an explicit MDP, a linear-Gaussian Kalman filter, reservoir and weighted stream sampling, and constant-memory Count-Min/HyperLogLog sketches.
- [Asset degradation state](https://docs.algenta.ai/optimization-and-simulation/degradation-state.md): Track an asset whose performance decays with usage age — a two-state filter with wear clamping, load correction and a cliff term for the non-linear end of life.
- [Forecasting and backtesting](https://docs.algenta.ai/time-series-and-statistics/forecasting.md): Forecast a series, measure the method out of sample, put an honest band around it, and make a hierarchy add up — with a refusal wherever the answer does not exist.
- [ES-RNN residual correction](https://docs.algenta.ai/time-series-and-statistics/ml-esrnn.md): ES-RNN-class residual correction over a frozen exponential-smoothing backbone — backbone selection, a per-series-adaptive window feature, and two correctors that consume it.
- [Time-series decomposition and queueing kernels](https://docs.algenta.ai/time-series-and-statistics/time-series-decomposition-and-queueing-kernels.md): Closed-form M/M/1, M/M/c, and Erlang B/C queueing formulas, plus additive time-series decomposition (trend, seasonal, residual) and a lightweight STL-like iterator -- two unrelated formula libraries s
- [Forecast scoring & Murphy decomposition](https://docs.algenta.ai/time-series-and-statistics/forecast-scoring.md): Score probabilistic forecasts honestly — Brier skill, calibration error, and the Murphy decomposition that separates being wrong from being overconfident.
- [Statistical inference](https://docs.algenta.ai/time-series-and-statistics/statistical-inference.md): p-values that carry the verdict on their own assumptions, sample sizes verified by re-evaluating power, reproducible resampling, and a critical value that is absent rather than clamped.
- [A/B testing and causal inference kernels](https://docs.algenta.ai/time-series-and-statistics/ab-testing-and-causal-inference-kernels.md): Two-arm experiment sizing and bandit selection, a compact single-formula causal-effects toolkit, classical decision theory under uncertainty, an ML-to-decision fusion layer, and five nonparametric hyp
- [Causal inference](https://docs.algenta.ai/time-series-and-statistics/causal-inference.md): Identification first — d-separation and the backdoor criterion on a declared graph — then IPW, matching, doubly-robust and panel estimation, and sensitivity analysis, refusing when not identified.
- [Survival analysis](https://docs.algenta.ai/time-series-and-statistics/survival-analysis.md): Time-to-event estimation that respects censoring — Kaplan-Meier with confidence bands, Nelson-Aalen, a kernel-smoothed hazard rate, Cox regression with Efron ties
- [Evaluation, survival, and psychometric kernels](https://docs.algenta.ai/time-series-and-statistics/ml-evaluation-survival-and-psychometric-kernels.md): Frozen hypothesis-test, survival-curve, operational-metrics, psychometric, and interval-arithmetic primitives that other kernel families call rather than reimplement.
- [Evaluation & calibration gates](https://docs.algenta.ai/time-series-and-statistics/evaluation-calibration.md): Brier with Murphy's decomposition, log-loss that says when it was clamped, calibration error that never travels without its binning, reproducible bootstrap intervals
- [Program execution & scoring](https://docs.algenta.ai/time-series-and-statistics/program-execution-scoring.md): Executes a resolved FinQA-style arithmetic program — ten operations, steps that reference earlier results by index — to the official evaluator's own rounding spec, bit-for-bit.
- [Sparse systems & graph analytics](https://docs.algenta.ai/data-and-retrieval/sparse-and-graph.md): Solve sparse linear systems and analyse graphs with compiled kernels on your own machine.
- [Retrieval & ranking](https://docs.algenta.ai/data-and-retrieval/retrieval.md): Search, fuse, rerank and explain results with compiled kernels on your own machine — and find out when a backend is missing instead of getting an empty list.
- [Retrieval fusion and signal kernels](https://docs.algenta.ai/data-and-retrieval/retrieval-fusion-and-signal-kernels.md): Reciprocal-rank fusion, hybrid blending, MMR reranking and rank-quality ratios over N ranked lists, plus the moving-average, EMA, IIR and spectral-adjacent filters that operate on one numeric series.
- [Retrieval, recommendation and ranking kernels](https://docs.algenta.ai/data-and-retrieval/retrieval-recommendation-and-ranking-kernels.md): RAG-specific fusion and quality scores, user-user collaborative filtering, ground-truth ranking metrics, vector-embedding arithmetic, and credit-style scoring formulas — five independent scalar librar
- [Data quality & feature hygiene](https://docs.algenta.ai/data-and-retrieval/data-quality.md): Expectations that name the failing row and column, a profile that keeps null, NaN and zero apart, duplicates under a key and an equality you declare
- [Geospatial indexing & joins](https://docs.algenta.ai/data-and-retrieval/geospatial.md): A spatial index with a measured selectivity, a three-valued point-in-polygon with no epsilon, coordinate frames that refuse a swapped latitude, and joins
- [Streaming sketches](https://docs.algenta.ai/data-and-retrieval/streaming-sketches.md): Distinct counts, frequencies and quantiles from summaries that fit in constant memory, merge across workers, survive a restart, and report absence rather than zero when there is no answer.
- [Windowed aggregates](https://docs.algenta.ai/data-and-retrieval/streaming-windows.md): Rolling minimum, maximum, range, mean and standard deviation over a sliding window, plus exponentially weighted mean and variance — with the real cost of each one charged before it runs.
- [Columnar queries](https://docs.algenta.ai/data-and-retrieval/columnar-queries.md): Filter, group, sort and rank tabular data with compiled kernels on your own machine.
- [Spreadsheet formula plane](https://docs.algenta.ai/data-and-retrieval/sheet-formula-plane.md): Recalculate real spreadsheet formulas — a tokenizer, dependency graph, cycle detection and 48 census-ranked functions, run as compiled kernels inside the runtime that ships with the SDK.
- [Text processing kernels](https://docs.algenta.ai/data-and-retrieval/text-processing-kernels.md): Fuzzy string distance, a hand-rolled regex engine, generic string utilities, word/sentence tokenization and a lexical column-name scorer, all zero-Python-import Mojo.
- [Text processing and tokenization kernels](https://docs.algenta.ai/data-and-retrieval/text-processing-and-tokenization-kernels.md): Corpus-linguistics readability and diversity statistics, a small from-scratch IR toolkit (TF-IDF, cosine, Jaccard), lexicon-based sentiment scoring, and the arithmetic primitives behind BPE/SentencePi
- [Text similarity and pattern matching kernels](https://docs.algenta.ai/data-and-retrieval/text-similarity-and-pattern-matching-kernels.md): A second, broader string-distance library, a minimal backtracking regex engine distinct from the corpus's other one, and vector-similarity formulas that operate on pre-reduced scalars rather than raw
- [Vector ranking kernels](https://docs.algenta.ai/data-and-retrieval/vector-ranking-kernels.md): SIMD Float32 dot-product and cosine-similarity primitives, a resident candidate-vector table, and the JSON-in/JSON-out column-name and repository-file rankers built on top of them.
- [Graph-based reasoning](https://docs.algenta.ai/data-and-retrieval/graph-based-reasoning-kernels.md): Scalar formulas for compiling an execution graph, scoring decisions walked over a graph under real constraints, and blending ML, simulation and symbolic rules into one reasoning step.
- [Data structures and algorithms kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/data-structures-and-algorithms-kernels.md): Binary heaps, a priority queue and circular buffer, a byte-level trie, six sorting algorithms, and sorted-list set operations -- pure-Mojo replacements for heapq, queue, pygtrie, sort, and set.
- [Bit-level and hashing kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/bit-level-and-hashing-kernels.md): Bit-twiddling primitives, non-cryptographic checksums and check digits, a counting Bloom filter, counter-based Philox4x64-10 random draws, and perceptual image hashing for near-duplicate detection.
- [Caching and persistence kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/caching-and-persistence-kernels.md): A TTL cache with hit-rate tracking, an in-memory table/query engine, a key-value store with snapshot/restore, and closed-form memory-sizing arithmetic for pools, allocators, and transformer training.
- [Encoding and cryptography kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/encoding-and-cryptography-kernels.md): Categorical data encoding, closed-form sizing arithmetic for block-cipher modes, classical-cryptanalysis statistics, and closed-form blockchain/consensus metrics — none of which perform an actual cryp
- [Networking and social graph kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/networking-and-social-graph-kernels.md): HTTP/URL protocol primitives, undirected and directed graph-metric arithmetic over adjacency lists, and the scalar cost formulas behind multi-GPU collective communication.
- [Reliability and observability kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/reliability-and-observability-kernels.md): Distributed-training checkpoint/recovery math, LLM-serving health metrics, a structured JSON logger, classical MTBF/Weibull reliability engineering, and perturbation-based sensitivity analysis -- five
- [Scheduling and rate-limiting kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/scheduling-and-rate-limiting-kernels.md): OS-course CPU scheduling algorithms and their metrics, stateless queue/backpressure/rate-limit scoring formulas, three independent token-bucket-family rate limiters, LLM-inference latency micro-optimi
- [Governance and workflow kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/governance-and-workflow-kernels.md): Petri net formalism primitives, real differential-privacy mechanisms, model/audit governance scoring, ESG carbon-accounting arithmetic, and the scalar primitives behind a DAG workflow orchestrator --
- [Utility and control primitive kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/utility-and-control-primitive-kernels.md): Package-dependency graph resolution, portfolio/Kelly-criterion allocation, statistical anomaly detection, backtracking constraint satisfaction, PID/control-theory step-response math, and numerical (no
- [System I/O and serialization kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/system-io-and-serialization-kernels.md): Environment-variable access, pure-string path and URL parsing, weighted/stratified random sampling, MessagePack binary serialization, and the histogram/dedup pair a richer, mostly-undispatchable conta
- [Streaming, structured output, and time kernels](https://docs.algenta.ai/systems-and-infrastructure-kernels/streaming-structured-output-and-time-kernels.md): Nine unrelated infrastructure primitives grouped only by being what is left over -- windowed stream aggregates, LLM structured-output scoring, time-zone/date/unit conversion, glob and template string
- [Image feature and quality kernels](https://docs.algenta.ai/vision-and-media-processing/image-feature-and-quality-kernels.md): Sobel edge/saliency features and color histograms, deterministic blur/exposure/entropy quality scores, pixel-level similarity metrics (MSE, PSNR, SSIM), and Otsu-based foreground region boxes over fla
- [Image transform and morphology kernels](https://docs.algenta.ai/vision-and-media-processing/image-transform-and-morphology-kernels.md): Binary morphology and connected components, grayscale convolution/filtering, box and bilinear downscaling with integral images, and deterministic RGB palette extraction, all on flat or nested Float64
- [Color science kernels](https://docs.algenta.ai/vision-and-media-processing/color-science-kernels.md): RGB/HSL/HSV color-space conversion, CIE colorimetry (delta-E, Lab-to-LCH, CCT, CRI, whiteness), and WCAG text-contrast accessibility checking, layered as three modules of increasing specificity.
- [Document vision and layout kernels](https://docs.algenta.ai/vision-and-media-processing/document-vision-kernels.md): Pixel-level document image processing -- Otsu and adaptive binarization, skew and scan-quality scoring, and projection-based segmentation into text lines, columns, density grids, and table-likeness sc
- [Video quality and accessibility kernels](https://docs.algenta.ai/vision-and-media-processing/video-quality-and-accessibility-kernels.md): Per-pixel worst-case WCAG contrast for caption overlays, dominant/mean color analysis, a composite no-reference frame-quality score, and clip-to-clip duplicate detection, all over flat row-major video
- [Video content and scene kernels](https://docs.algenta.ai/vision-and-media-processing/video-content-and-scene-kernels.md): Frame-level video primitives -- scene-change scoring, motion-region extraction, storyboard/keyframe selection, and the routing gate that decides how much of a clip is worth a model call, mirroring the
- [Ranging and remote-sensing kernels](https://docs.algenta.ai/vision-and-media-processing/ranging-and-remote-sensing-kernels.md): LiDAR point-cloud geometry and vegetation-return metrics, LiDAR-derived terrain/canopy mapping and survey QA, and the core radar range/Doppler/link-budget equations.
- [Solid-Earth geophysics](https://docs.algenta.ai/earth-and-planetary-sciences/solid-earth-geophysics-kernels.md): WGS84 geodesy and coordinate conversion, gravity/magnetic/seismic-velocity survey formulas, earthquake magnitude and travel-time relations, and simplified volcanic-eruption dynamics.
- [Surface processes & geomorphology](https://docs.algenta.ai/earth-and-planetary-sciences/surface-processes-and-geomorphology-kernels.md): Landform evolution and fluvial morphometrics, soil/rock mechanics for construction, soil formation and classification science, and sediment transport/grain-size statistics, across four independent sca
- [Hydrology & water systems](https://docs.algenta.ai/earth-and-planetary-sciences/hydrology-and-water-systems-kernels.md): Surface hydraulics, groundwater flow, lake physics and chemistry, ocean dynamics, storm-water meteorology and the vegetation-water interface across seven independent kernel modules.
- [Historical geology & paleoclimate](https://docs.algenta.ai/earth-and-planetary-sciences/historical-geology-and-paleoclimate-kernels.md): Radiometric dating, rock-magnetic paleolatitude, sediment-layer age modeling, and ice-sheet mass balance across four independent Earth-history kernel modules.
- [Speleology, mineralogy & geochemistry](https://docs.algenta.ai/earth-and-planetary-sciences/speleology-mineralogy-and-geochemistry-kernels.md): Cave hydrology, speleothem dating, and cave-radio physics; mineral identification and crystal chemistry; X-ray diffraction and unit-cell geometry; and the trace-element, isotope, and elemental-cycling
- [Cartography, planetary & cosmic sciences](https://docs.algenta.ai/earth-and-planetary-sciences/cartography-and-planetary-cosmic-sciences-kernels.md): Map-scale and projection arithmetic for cartography, Julian-date and orbital-mechanics formulas for astronomy, Hubble-law cosmology and blackbody radiation laws, and habitability/biosignature scoring
- [Atmospheric chemistry, climate & meteorology](https://docs.algenta.ai/earth-and-planetary-sciences/atmospheric-chemistry-and-climate-kernels.md): Gas-phase photochemistry and pollutant-lifetime formulas, plus two independently-implemented single-point weather-index toolkits that compute instantaneous psychrometric conditions rather than long-te
- [Earth, electrical, and neurophysics kernels](https://docs.algenta.ai/earth-and-planetary-sciences/earth-electrical-and-neurophysics-kernels.md): Spherical-Earth navigation math, textbook DC/AC circuit-theory formulas, and the biophysics of neuron membranes and action potentials, across three unrelated small Float64 kernel modules.
- [Acoustics, optics & radiative materials](https://docs.algenta.ai/physical-and-materials-science/acoustics-optics-and-radiative-materials-kernels.md): Five independent physics-formula modules for sound and light — decibel levels and room acoustics, medical/industrial ultrasound imaging physics, blackbody and photon radiometry, fiber/optical photonic
- [Structural & construction materials](https://docs.algenta.ai/physical-and-materials-science/structural-and-construction-materials-kernels.md): Adhesive bond strength and cure kinetics, fiber-composite mechanics, concrete mix design and curing, ceramic fracture statistics, polymer chain/molecular-weight science, and rubber/elastomer testing,
- [Continuum and structural mechanics kernels](https://docs.algenta.ai/physical-and-materials-science/continuum-and-structural-mechanics-kernels.md): Classical point mechanics, beam/column structural analysis, Newtonian pipe-and-body fluid dynamics, non-Newtonian and viscoelastic rheology, and soil engineering split across its geotechnical and agro
- [Corrosion & electrochemical materials](https://docs.algenta.ai/physical-and-materials-science/corrosion-and-electrochemical-materials-kernels.md): Corrosion rate and cathodic protection, electrochemical cell and battery equations, classical electrostatics and magnetism, metallurgical strengthening/hardenability, and semiconductor device physics
- [Thermal, cryogenic & energy materials](https://docs.algenta.ai/physical-and-materials-science/thermal-and-energy-materials-kernels.md): Cryogenic and permafrost thermal physics, the four classical thermodynamic relations, battery electrochemistry, and solar/wind/photovoltaic energy-yield engineering across seven independent kernel mod
- [Nanoscale, plasma & surface-contact engineering](https://docs.algenta.ai/physical-and-materials-science/advanced-nano-and-plasma-materials-kernels.md): Nanoscale surface/transport physics, plasma and fusion parameters, dry and lubricated contact mechanics with wear-corrosion coupling, and antenna/RF link-budget equations across five independent kerne
- [Process & organic-materials science](https://docs.algenta.ai/physical-and-materials-science/process-and-organic-materials-science-kernels.md): Small scalar-formula libraries for eight process and organic-materials domains — brewing, bioprocess fermentation, food science, fragrance chemistry, pulp and paper, timber, textiles, and combustion.
- [Materials analysis and spectroscopy kernels](https://docs.algenta.ai/physical-and-materials-science/materials-analysis-and-spectroscopy-kernels.md): Generic mechanics-of-materials formulas, chromatographic separation science, optical/analytical spectroscopy, treatment-plant water chemistry, and structural timber engineering across five independent
- [Clinical specialty kernels](https://docs.algenta.ai/medical-and-health-sciences/clinical-specialty-kernels.md): Textbook clinical formulas for seven specialties -- cardiac hemodynamics, dental/oral-health indices, hormone physiology, kidney function, eye optics, orthodontic biomechanics, and lung function -- as
- [Immunology, hematology, and virology kernels](https://docs.algenta.ai/medical-and-health-sciences/immunology-hematology-and-virology-kernels.md): Antibody/immune-response, blood-cell-index, viral-kinetics, cell-biology, and flow-cytometry formulas across five independent clinical and lab-science modules.
- [Pharmacology and toxicology kernels](https://docs.algenta.ai/medical-and-health-sciences/pharmacology-and-toxicology-kernels.md): One-compartment drug pharmacokinetics and dose-response pharmacology, CNS-specific receptor math, the parallel toxicokinetics/toxicology pair for toxic and environmental substances, and radiation dosi
- [Epidemiology and disease dynamics kernels](https://docs.algenta.ai/medical-and-health-sciences/epidemiology-and-disease-dynamics-kernels.md): Compartmental disease-spread models and outbreak metrics, the risk-factor and diagnostic-test statistics used to argue a cause, and molecular assay quantification for gene regulation that never touche
- [Biomedical, biomechanics, and psychoacoustics kernels](https://docs.algenta.ai/medical-and-health-sciences/biomedical-engineering-kernels.md): Bedside cardiovascular/renal/respiratory/pharmacokinetic formulas, gait and joint-mechanics equations, and human sound-perception models — three unrelated small physics/physiology libraries, not one "
- [Nutrition and veterinary science kernels](https://docs.algenta.ai/medical-and-health-sciences/nutrition-and-veterinary-science-kernels.md): Human energy-balance and body-composition formulas (BMI, Mifflin-St Jeor BMR, TDEE, macros) alongside weight-based veterinary dosing, fluid therapy, and per-species reference constants for dogs, cats,
- [Agricultural production kernels](https://docs.algenta.ai/biology-and-agricultural-sciences/agricultural-production-kernels.md): Five flat-scalar modules for crop and plant production math -- field agriculture, agronomic crop/soil science, horticultural plant physiology, hydroponic nutrient/climate control, and greenhouse struc
- [Specialty agriculture & aquaculture kernels](https://docs.algenta.ai/biology-and-agricultural-sciences/specialty-agriculture-and-aquaculture-kernels.md): Per-hive, per-tank, per-stock, per-hectare, and per-trap arithmetic for six specialty production sciences -- beekeeping, aquatic farming, wild-fisheries stock assessment, wine-grape growing and cellar
- [Molecular & cell biology kernels](https://docs.algenta.ai/biology-and-agricultural-sciences/molecular-and-cell-biology-kernels.md): Sequence-level bioinformatics, genome-scale statistics, mass-spectrometry protein and metabolite quantification, allele-frequency population genetics, and plant secondary-metabolite assay formulas acr
- [Ecology & organismal biology kernels](https://docs.algenta.ai/biology-and-agricultural-sciences/ecology-and-organismal-biology-kernels.md): Population-dynamics and community-diversity formulas for general ecology, plus five taxon-focused libraries -- ecophysiology, entomology, microbiology, mycology, and bryology -- covering organism-envi
- [Paleobiology & archaeobotany kernels](https://docs.algenta.ai/biology-and-agricultural-sciences/paleobiology-and-archaeobotany-kernels.md): Tree-ring dating and climate-signal statistics, forest-mensuration biometrics, quantitative pollen/spore analysis, and fossil preservation and decay-process formulas across four independent historical
- [Aerospace & autonomous systems kernels](https://docs.algenta.ai/engineering-and-industrial-systems/aerospace-and-autonomous-systems-kernels.md): Orbital and rocket mechanics plus atmospheric aerodynamics (aerospace), aircraft-specific performance (aviation), UAV power/range/imaging (drones), self-driving-car safety physics (autonomous\_vehicles
- [Building & facility systems kernels](https://docs.algenta.ai/engineering-and-industrial-systems/building-and-facility-systems-kernels.md): Vertical-transport and building-facility engineering -- elevator dispatch/sizing, fire-safety egress and suppression, HVAC heat load, plumbing hydraulics, refrigeration cycles, wastewater treatment de
- [Industrial fluid power & energy systems kernels](https://docs.algenta.ai/engineering-and-industrial-systems/industrial-fluid-power-and-energy-systems-kernels.md): Hydraulic (pressurized-liquid) power transmission, pneumatic (compressed-air) power transmission, and grid-scale AC power engineering — three independent kernel modules, none of which overlap with the
- [Transportation & hazard-safety kernels](https://docs.algenta.ai/engineering-and-industrial-systems/transportation-and-hazard-safety-kernels.md): Traffic-signal timing and level-of-service formulas, physical oceanography, mine-planning economics with blasting and ventilation arithmetic, and nuclear/reactor decay physics — four independent engin
- [Manufacturing & craft engineering kernels](https://docs.algenta.ai/engineering-and-industrial-systems/manufacturing-and-craft-engineering-kernels.md): Joint-strength and heat-control physics for arc welding, mechanical-watch gear trains and escapement timing, press production math for printing, corrugated-box and container engineering, and textile-c
- [Accounting, actuarial, insurance & pension kernels](https://docs.algenta.ai/finance-business-and-economics/accounting-and-actuarial-kernels.md): Depreciation, financial ratios, and time-value-of-money accounting; actuarial mortality, life-expectancy, and reserve math; insurance premium ratemaking and loss-ratio formulas; and pension funding, v
- [Auction & market mechanism kernels](https://docs.algenta.ai/finance-business-and-economics/auction-and-market-mechanism-kernels.md): Auction revenue and payment rules (Vickrey/English/Dutch/VCG), closed-form behavioral-economics biases, and voting rules computed directly from raw ballots.
- [Macroeconomic indicators & currency kernels](https://docs.algenta.ai/finance-business-and-economics/macroeconomic-indicators-and-currency-kernels.md): Time-value-of-money and loan math, corporate/investment finance ratios, national-economy indicators (GDP, inflation, unemployment), and human population statistics -- four independent formula librarie
- [Risk, supply-chain & logistics kernels](https://docs.algenta.ai/finance-business-and-economics/risk-supply-chain-and-logistics-kernels.md): Portfolio/returns-series risk metrics (VaR, Sharpe, drawdown) alongside five independent operations-research modules for demand forecasting, EOQ/safety-stock, procurement economics, logistics KPIs, an
- [Business operations & real-estate kernels](https://docs.algenta.ai/finance-business-and-economics/business-operations-and-real-estate-kernels.md): Queueing/inventory/LP-scoring formulas, earned-value and PERT project math, real-estate underwriting ratios and mortgage amortization, and telecom link-budget/traffic-engineering formulas -- four inde
- [Collectibles, gaming & sports analytics kernels](https://docs.algenta.ai/finance-business-and-economics/collectibles-gaming-and-sports-analytics-kernels.md): Museum conservation and collections-management arithmetic, coin valuation and grading, lottery combinatorics and expected-value math, and sports performance metrics, across four independent scalar-for
- [Humanities and social science kernels](https://docs.algenta.ai/humanities-and-social-sciences/humanities-and-social-science-kernels.md): Radiocarbon/stratigraphic dating and artifact statistics for archaeology, NIOSH/Fitts's-law/Hick's-law human-factors formulas for ergonomics, time-of-death/toxicology/ballistics calculations for foren
- [How the engine works](https://docs.algenta.ai/concepts/engine.md): How Algenta integrates deterministic compiled compute, validates every request, and fails closed without exposing runtime lifecycle to users.
- [The governed decision lifecycle](https://docs.algenta.ai/concepts/decision-service.md): How Algenta turns enterprise data and operating models into governed, versioned, reproducible decisions instead of one-off chat answers
- [Governed data & query](https://docs.algenta.ai/concepts/governed-data.md): Governed data is data Algenta queries under a typed, schema-validated, deterministic, auditable contract — sources, datasets, connectors, and the resolve to query flow.
- [The capability plane](https://docs.algenta.ai/concepts/capability-plane.md): The unified router that turns an objective into the right provider and skill, then executes it — discovery, routing, execution, adapters, and fail-closed execution ownership.
- [Simulation models & distributions](https://docs.algenta.ai/concepts/simulation-models.md): The simulation algorithms and input distributions Algenta runs — auto vs expert mode, when to pick each engine, and how a seed makes every run reproducible.
- [Decision memory](https://docs.algenta.ai/concepts/decision-memory.md): Decision memory is the persisted log of decisions, the real-world outcomes recorded against them, and the review loop that lets the engine learn which decision types are over- or under-confident.
- [Execution ownership & fail-closed](https://docs.algenta.ai/concepts/execution-ownership.md): Who runs a capability — algenta\_managed (the engine executes) vs client\_managed (your adapter executes) — and why an unresolved or unsupported owner fails closed.
- [Deployment modes & privacy](https://docs.algenta.ai/concepts/deployment-modes.md): How ALGENTA\_DEPLOYMENT\_MODE selects a privacy profile — cloud, telemetry, metering, and egress defaults — and how the engine fails closed on outbound network in air-gapped runs.
- [Make your first query](https://docs.algenta.ai/guides/first-query.md): Run your first governed query against a registered dataset and read the deterministic result, using HTTP, the de CLI, or the Python SDK.
- [Group data locally with Query.group\_by()](https://docs.algenta.ai/guides/local-query-groupby.md): Build a columnar query with algenta.Query and group your own data locally with group\_by() — the same composed engine the dataframe group-by benchmark measures.
- [Run a decision](https://docs.algenta.ai/guides/run-a-decision.md): Turn a scenario into a recommended action — call simulate or recommend and read the decision envelope (recommended action, expected value, confidence, percentiles).
- [Drive a governed decision](https://docs.algenta.ai/guides/governed-decisions.md): Frame a decision case, run the governed analysis, inspect the evidence, then propose, approve, and execute an action plan through the host-independent Decision API
- [Renewal Capacity Allocation pack](https://docs.algenta.ai/guides/renewal-capacity-allocation.md): Allocate limited customer-success hours across upcoming renewals to maximize model-estimated retained ARR, modeling intervention effects explicitly under hard constraints.
- [Simulate a scenario](https://docs.algenta.ai/guides/simulate.md): Build a full simulation request — auto vs expert mode, choose a simulation\_model, declare distributions and params, pin a seed, and read the percentiles back.
- [Kernels on Hugging Face](https://docs.algenta.ai/guides/kernels-on-hugging-face.md): Load Algenta compute kernels on Apple silicon straight from the Hugging Face Kernel Hub with the kernels library.
- [Run a what-if sweep](https://docs.algenta.ai/guides/what-if.md): Sweep one simulation variable across a range to see how outcomes move, and stream a long simulation live over Server-Sent Events.
- [Recommend, score & compare](https://docs.algenta.ai/guides/recommend-and-score.md): Compare named actions and get a ranked recommendation, score a single scenario with custom weights, run many simulations at once, and compare scenarios side by side.
- [Query governed data](https://docs.algenta.ai/guides/governed-query.md): Run a deterministic governed query end to end — resolve structured intent into an exact plan, execute it with typed filters, and read plan\_hash and validated.
- [Log decisions & record outcomes](https://docs.algenta.ai/guides/decision-memory.md): Log a decision through the engine, record its real-world outcome later, and review the closed loop so the engine learns which decision types are over or under confident.
- [Drive decisions through the API](https://docs.algenta.ai/guides/decision-memory-api.md): Drive the full Algenta decisions API — plan, log, execute, and delete decisions — and call the one-line product endpoints for decision, agent, optimize, retrieve, and forecast.
- [Configure the execution policy](https://docs.algenta.ai/guides/execution-policy.md): Read and update the Algenta execution policy — the confidence, risk, calibration, and idempotency gates that guard autonomous decision execution — and inspect calibration.
- [Connect a data source](https://docs.algenta.ai/guides/connect-data.md): Register an external source as a saved connector, test it, browse its schema, and onboard a dataset so the engine can query it under a typed contract.
- [Map a dataset instantly](https://docs.algenta.ai/guides/instant-map.md): Drop in CSV, JSON, or connected tables and get back a unified schema with every field mapped, join keys detected, and confidence for each match — no schema, no ETL.
- [Route & execute capabilities](https://docs.algenta.ai/guides/capabilities.md): Route an objective to the right capability and execute it through the capability plane, using HTTP, the de CLI, or the Python SDK, with execution ownership made explicit.
- [Register capability providers & bindings](https://docs.algenta.ai/guides/capability-bindings.md): Register capability providers, create and test bindings, discover and authorize capabilities, and enable skills — the setup layer behind the Algenta capability plane.
- [Connect an external MCP server](https://docs.algenta.ai/guides/external-mcp.md): Register an external MCP server with Algenta — discover its tools into the unified capability catalog and let the engine call them on your agent's behalf, all under the egress policy.
- [Run an agent](https://docs.algenta.ai/guides/agent-runs.md): Start a deterministic, replayable agent run through the Algenta Agent Runtime, follow its lifecycle, stream its events, and retrieve its checkpoints and proof artifacts.
- [Run an async job](https://docs.algenta.ai/guides/async-jobs.md): Submit a long-running simulation as an async job, poll its status by id, fetch the result when it completes, and cancel a queued or running job.
- [Triage a repository](https://docs.algenta.ai/guides/repository-intelligence.md): Triage a code repository with the Algenta engine — snapshot it, reduce it to a bounded evidence bundle, and read back the findings over the HTTP API
- [Fix a repository](https://docs.algenta.ai/guides/repository-fix.md): Turn a triage evidence bundle into a verified code fix — generate a decision plan, simulate its risk gate, and apply it as a patch, a local branch, or a pull request.
- [Prove a security finding](https://docs.algenta.ai/guides/security-analysis.md): Prove reachability for scanner findings against a repository snapshot, triage by classification, and re-prove that a fix closes the finding after you patch it.
- [Register a trigger](https://docs.algenta.ai/guides/triggers.md): Register a threshold condition on a data source, fire it to run an attached simulation, pause or resume it, and delete it — with optional webhook and auto-execution dispatch.
- [Webhooks](https://docs.algenta.ai/guides/webhooks.md): Receive Algenta events by webhook — async job callbacks, trigger notifications, and decision-execution dispatch — verify delivery with a test POST, and receive inbound GitHub App events.
- [Bring your own key (BYOK)](https://docs.algenta.ai/guides/byok.md): Supply your own LLM-provider or engine credential. Algenta stores it encrypted at rest, never returns the key material, and uses it only to serve your own org's calls.
- [Device binding](https://docs.algenta.ai/guides/device-binding.md): Bind an Algenta offline license to a machine identity so it cannot be copied, print the device id, verify a bound license, and fix a device\_identity\_mismatch.
- [Run air-gapped](https://docs.algenta.ai/guides/air-gapped.md): Run Algenta fully air-gapped with no outbound network — install and verify an offline RS256 license, prove the egress policy fails closed, and set the engine for no-LLM or BYOK operation.
- [Use Algenta in CI/CD](https://docs.algenta.ai/guides/ci-cd.md): Run Algenta decisions, simulations, and repository triage from CI/CD with the de CLI — authenticate from a secret, emit JSON, and gate the pipeline on exit codes.
- [Run local LLM decode](https://docs.algenta.ai/guides/run-qwen3-decode.md): Install Algenta, get a supported checkpoint, and run greedy decode on your own Apple Silicon GPU — requirements, limitations, and a complete working example. Currently tested on Qwen3 models.
- [Model-agnostic LLM inference](https://docs.algenta.ai/guides/run-qwen3-family-inference.md): Run any supported checkpoint on Apple Silicon through Algenta's model-agnostic SDK path — supported shapes, long context, and the quantized load cache. Currently verified on Qwen3 models.
- [LLM models & providers](https://docs.algenta.ai/llm-api/models.md): The Algenta LLM model catalog — GET /v1/models, the provider.\<backend>:\<model> id scheme, the ten configurable backends, dynamic model discovery, and per-backend BYOK.
- [Run chat completions](https://docs.algenta.ai/llm-api/completions.md): Run chat completions, the unified response envelope, tokenize, and count\_tokens over the Algenta LLM API — with streaming, from curl, Python, TypeScript, or the CLI.
- [Generate embeddings](https://docs.algenta.ai/llm-api/embeddings.md): Generate text embeddings, score two vectors with a similarity metric, and rerank candidate documents over the Algenta LLM API — from curl, Python, TypeScript, or the CLI.
- [Python SDK](https://docs.algenta.ai/sdks/python.md): Use Algenta from Python — install algenta-sdk, authenticate, and call the engine with a typed client. Includes runnable examples.
- [Python client reference](https://docs.algenta.ai/sdks/python/overview.md): Install algenta-sdk, construct the sync or async AlgentaClient, configure retries and timeouts, handle typed errors, and jump to the Python method reference by domain.
- [Simulations & queries](https://docs.algenta.ai/sdks/python/simulations-and-queries.md): Python SDK reference for the simulation and query methods — simulate, resolve, query, query\_batch, verify, recommend, score, compare, and plan\_decision.
- [Decisions & products](https://docs.algenta.ai/sdks/python/decisions-and-products.md): Python SDK reference for decision memory and product intelligence — log\_decision, record\_outcome, execute\_decision, plus product\_decision, product\_optimize, and product\_forecast.
- [Connectors & datasets](https://docs.algenta.ai/sdks/python/connectors-and-data.md): Python SDK reference for connectors and datasets — create, test, browse, and update connectors, then onboard, list, summarize, refresh, and delete datasets.
- [Repository intelligence](https://docs.algenta.ai/sdks/python/repositories.md): Python SDK reference for repository intelligence — capabilities, snapshots, triage, decision plans, graph queries, simulate, and apply.
- [LLM & embeddings](https://docs.algenta.ai/sdks/python/llm.md): Python SDK reference for the governed LLM surface — list\_models, tokenize, count\_tokens, chat\_completions, responses, embeddings, embedding\_similarity, and rerank.
- [Agent runs](https://docs.algenta.ai/sdks/python/agent-runs.md): Python SDK reference for agent runs — launch, inspect, stream, pause/approve/resume, replay and fork agentic runs, with every AlgentaClient method and its endpoint.
- [Jobs, triggers & capabilities](https://docs.algenta.ai/sdks/python/jobs-triggers-capabilities.md): Python SDK reference for async jobs, deployments, triggers, data sources, and the capability plane — every AlgentaClient method with its endpoint and return type.
- [Account & control plane](https://docs.algenta.ai/sdks/python/account-and-control-plane.md): Python SDK reference for identity, usage, API keys, teams, billing, devices, audit logs, execution policy, and the platform contract — every AlgentaClient method and endpoint.
- [TypeScript SDK](https://docs.algenta.ai/sdks/typescript.md): Use Algenta from TypeScript or JavaScript — install the SDK, authenticate, and call the engine with a typed client. Includes runnable examples.
- [TypeScript client reference](https://docs.algenta.ai/sdks/typescript/overview.md): Install algenta-sdk, construct an AlgentaClient, and learn the promise-based method surface and the typed error hierarchy before you call the engine.
- [Simulations & queries](https://docs.algenta.ai/sdks/typescript/simulations-and-queries.md): TypeScript SDK methods for Monte Carlo simulation, decision planning, recommendation, scoring, and the governed query family, each with its HTTP route.
- [Decisions & products](https://docs.algenta.ai/sdks/typescript/decisions-and-products.md): TypeScript SDK methods for decision memory (log, list, outcome, execute, delete) and the high-level product endpoints for decision, optimize, retrieve, forecast, and agent runs.
- [LLM & model discovery](https://docs.algenta.ai/sdks/typescript/llm.md): TypeScript SDK methods for the governed LLM surface — chat, responses, embeddings, rerank, tokenization, model artifacts — plus model catalog and runtime discovery.
- [Connectors & data](https://docs.algenta.ai/sdks/typescript/connectors-and-data.md): TypeScript SDK methods for connectors, datasets, source registration, and repository intelligence — create, test, browse, onboard, snapshot, triage, and apply.
- [Repository intelligence](https://docs.algenta.ai/sdks/typescript/repositories.md): TypeScript SDK reference for repository intelligence — snapshot a repo, triage an issue into ranked evidence, build a decision plan, query the symbol graph, simulate, and apply a fix.
- [Agent runs](https://docs.algenta.ai/sdks/typescript/agent-runs.md): TypeScript SDK reference for agent runs — launch, stream, approve, fork, and replay runs, and query their events, checkpoints, mission events, and telemetry.
- [Jobs, triggers & capabilities](https://docs.algenta.ai/sdks/typescript/jobs-triggers-capabilities.md): TypeScript SDK reference for async jobs, webhook triggers, and the capability plane — submit and poll jobs, register triggers, and route, execute, and bind capabilities.
- [Account & control plane](https://docs.algenta.ai/sdks/typescript/account-and-control-plane.md): TypeScript SDK reference for the account and control plane — identity, usage, limits, billing, API keys, team, devices, audit logs, execution policy, deployments, health, and version.
- [MCP server](https://docs.algenta.ai/sdks/mcp.md): Connect Algenta to MCP-compatible tools like Cursor and Claude Desktop. Configure the MCP server, authenticate, and verify the tools appear.
- [Install from the MCP Registry](https://docs.algenta.ai/sdks/mcp-registry.md): Install the Algenta MCP server from the Official MCP Registry — io.github.thyn-ai/algenta — into Cursor, Claude Desktop, Claude Code, or Codex, passing your ALGENTA\_API\_KEY.
- [Tool discovery](https://docs.algenta.ai/sdks/tool-discovery.md): How an AI agent or MCP client discovers Algenta's tools — list the gated tool catalog, read the machine-readable contract, and check the runtime manifest before planning.
- [Testing with Algenta](https://docs.algenta.ai/sdks/testing.md): Test Algenta-backed code with the engine's determinism — assert stable decision\_hash values, replay recorded runs, mock the HTTP transport, and gate live tests on an API key.
- [CLI reference](https://docs.algenta.ai/sdks/cli.md): Install and use the supported Algenta command-line surfaces without managing a separate compute runtime.
- [MCP tools overview](https://docs.algenta.ai/mcp-tools/overview.md): The Algenta MCP server exposes the engine as Model Context Protocol tools for Cursor, Claude, and other hosts. Browse the tool catalog by domain.
- [Data & ingest tools](https://docs.algenta.ai/mcp-tools/data.md): MCP reference for Algenta data tools — onboard and train datasets, run the high-level connection flow, register sources, and ingest tables to a simulation.
- [Connector tools](https://docs.algenta.ai/mcp-tools/connectors.md): MCP reference for Algenta connector tools — create, inspect, test, browse, update, and delete saved data connectors, plus preview inline definitions.
- [Query tools](https://docs.algenta.ai/mcp-tools/query.md): MCP reference for Algenta query tools — governed exact queries, multi-metric batches, and constrained read-only SQL reports over your authorized datasets.
- [Repository intelligence tools](https://docs.algenta.ai/mcp-tools/repository-intelligence.md): MCP reference for Algenta's repository-intelligence tools — snapshot, triage, plan, graph-query, simulate, and apply a code fix from any MCP client.
- [LLM utility tools](https://docs.algenta.ai/mcp-tools/llm.md): MCP reference for Algenta's deterministic utility-model tools — list models, tokenize, count tokens, chat, responses, embeddings, similarity, and rerank.
- [Capability plane tools](https://docs.algenta.ai/mcp-tools/capability-plane.md): MCP reference for Algenta's capability plane — list providers and bindings, discover and test them, route an objective, execute a capability, and manage skills.
- [Decision tools](https://docs.algenta.ai/mcp-tools/decisions.md): MCP reference for Algenta's decision tools — simulate, recommend, score, compare, the product helpers, and the decision-memory log/outcome loop.
- [Async jobs & triggers tools](https://docs.algenta.ai/mcp-tools/jobs-and-triggers.md): MCP reference for async simulation jobs and real-time triggers — submit and poll long-running jobs, test webhooks, and register conditions that auto-run simulations.
- [Agent run tools](https://docs.algenta.ai/mcp-tools/agent-runs.md): MCP reference for the Algenta agent-run lifecycle — create runs, read events, checkpoints, mission events and telemetry, and resume, cancel, or approve a run.
- [Control plane tools](https://docs.algenta.ai/mcp-tools/control-plane.md): MCP reference for deployments, team members, devices, audit logs, execution policy, billing, and managed-runtime compatibility.
- [Account & meta tools](https://docs.algenta.ai/mcp-tools/account-and-meta.md): MCP reference for account, identity, API keys, usage and runtime discovery — the get\_contract, get\_me, key-management, run-history and analytics tools.
- [Algenta for Microsoft Cloud](https://docs.algenta.ai/microsoft-and-azure/overview.md): Turn enterprise data and operating models into governed decisions across Microsoft 365, powered by a decision runtime in your own Azure subscription
- [Decision Agent in Copilot & Teams](https://docs.algenta.ai/microsoft-and-azure/copilot-teams.md): A declarative Microsoft 365 Copilot and Teams agent that connects to your Algenta runtime over the Model Context Protocol and turns a conversation into a computed decision
- [Excel Decision Add-in](https://docs.algenta.ai/microsoft-and-azure/excel.md): Run a governed Decision Pack from an Excel task pane over your own decision gateway — selection-only, preview-first, results to new sheets
- [Sign in with Microsoft Entra](https://docs.algenta.ai/microsoft-and-azure/entra-sso.md): Authenticate employees and service identities to the Decision API with audience-scoped Microsoft Entra access tokens, then map each principal to an organization and workspace
- [Deploy the decision gateway to your Azure](https://docs.algenta.ai/microsoft-and-azure/gateway.md): Deploy the Algenta gateway into your own Azure subscription with a guided Bicep template that keeps the runtime and data plane customer-hosted with cloud features off
- [Sign in with a generic OIDC provider](https://docs.algenta.ai/identity-providers/generic-oidc.md): Authenticate users and service clients to the Decision API with any standards-compliant OIDC provider — Okta, Auth0, Keycloak, or your own — then map each principal to an organization and workspace
- [Self-hosting](https://docs.algenta.ai/deploy-and-operate/self-hosting.md): Deploy and operate the Algenta engine on your own infrastructure — configure it, run it, and verify a self-hosted instance is healthy.
- [Configuration reference](https://docs.algenta.ai/deploy-and-operate/configuration.md): Every Algenta configuration environment variable in one place, with its default and any aliases — complete and current.
- [Monitoring & observability](https://docs.algenta.ai/deploy-and-operate/monitoring.md): Monitor a self-hosted Algenta deployment — scrape the protected /metrics endpoint, read the engine health signals, tail the structlog JSON logs, all inside your own network.
- [Encryption at rest (BYOK)](https://docs.algenta.ai/deploy-and-operate/byok-kms.md): Encrypt Algenta secrets at rest with your own AES-256-GCM key supplied through ALGENTA\_CONFIG\_ENCRYPTION\_KEY.
- [Device key store](https://docs.algenta.ai/deploy-and-operate/device-key-store.md): Choose where Algenta stores the per-device private key — Apple Secure Enclave, a Linux TPM, the OS keyring, or an encrypted file — and confirm the runtime is healthy with runtime doctor.
- [Audit & tighten egress](https://docs.algenta.ai/deploy-and-operate/privacy-egress.md): Read the engine's live egress policy, tighten it with deployment-mode settings, generate a 24-hour privacy report, and audit the redacted egress decision log.
- [Teams & roles (RBAC)](https://docs.algenta.ai/deploy-and-operate/teams-rbac.md): Invite team members and set roles, understand the owner/admin/member/viewer permission matrix, and read the org audit log and its content-hashed artifacts.
- [Deployments & nodes](https://docs.algenta.ai/deploy-and-operate/deployments-and-nodes.md): Provision an isolated cloud deployment across AWS, Azure, or GCP, track its status and monthly cost, and register self-hosted worker nodes that heartbeat into the engine.
- [Licensing & editions](https://docs.algenta.ai/deploy-and-operate/licensing.md): How Algenta licensing works for self-hosted and air-gapped use — what a license grants, how to install and verify it offline, the editions, and the open-core terms.
- [Usage & limits](https://docs.algenta.ai/deploy-and-operate/billing.md): Inspect unlimited execution policy and local governance records. Algenta is licensed per device, never per decision.
- [Trust overview](https://docs.algenta.ai/trust-and-security/overview.md): Who is responsible for what across Algenta's deployment modes, and exactly which parties ever see your data.
- [Governed execution](https://docs.algenta.ai/trust-and-security/governed-execution.md): The control statements behind every plan approval and execution — named denial codes, single-use approval binding, versioned receipts.
- [Data boundaries](https://docs.algenta.ai/trust-and-security/data-boundaries.md): Every outbound network call the engine can make is classified into one of five categories and gated by deployment mode — plus what the SSRF guard blocks unconditionally.
- [Audit and evidence](https://docs.algenta.ai/trust-and-security/audit-and-evidence.md): The audit log export carries a content hash you re-derive yourself — evidence you verify, not evidence you take on trust.
- [Identity and access](https://docs.algenta.ai/trust-and-security/identity-and-access.md): Four separate identity mechanisms, each scoped to one surface, none falling through to another.
- [Capability status](https://docs.algenta.ai/trust-and-security/capability-status.md): Every protocol surface and integration, labeled Available now, Validated integration, or In conformance testing — checked against the current code and tests, not restated from an announcement.
- [Protocol compatibility](https://docs.algenta.ai/trust-and-security/protocol-compatibility.md): MCP version negotiation, the Responses and Chat Completions surfaces, and the 12-scenario conformance suite's real results — 9 passed, 0 failed, 3 blocked with named reasons.
- [Architecture and trust boundaries](https://docs.algenta.ai/trust-and-security/architecture.md): Where the engine, your model, and your MCP clients run relative to each other — and the one boundary every deployment mode shares.
- [Evidence walkthrough](https://docs.algenta.ai/trust-and-security/evidence-walkthrough.md): The versioned execution-receipt envelope, the audited webhook-dispatch path, and the hash-verified audit export — three real artifacts, and an honest note on how they connect today.
- [Security questionnaire](https://docs.algenta.ai/trust-and-security/security-questionnaire.md): Answers to the questions a vendor security review actually asks — each one checked against this repository's code and tests, with the gaps stated as plainly as the controls.
- [Supply chain](https://docs.algenta.ai/trust-and-security/supply-chain.md): What's verifiable about an Algenta release today — SBOMs and a vulnerability scan per image — and what isn't yet, stated as plainly as what is.
- [Threat model](https://docs.algenta.ai/trust-and-security/threat-model.md): A public summary of what this engine defends against and how — asset by asset, control by control — without exploit-level detail that would only help an attacker.
- [Support policy](https://docs.algenta.ai/trust-and-security/support-policy.md): What support and long-term-support commitments exist today, and what is explicitly pending an owner decision rather than invented ahead of it.
- [API overview](https://docs.algenta.ai/http-api/overview.md): The Algenta HTTP API — base URL, authentication, conventions, and where to find the full reference and the live explorer.
- [API endpoint reference](https://docs.algenta.ai/http-api/reference.md): The complete Algenta HTTP API endpoint reference — every operation grouped by area, with method, path, and summary, generated from the OpenAPI spec.
- [Streaming (SSE)](https://docs.algenta.ai/http-api/streaming.md): The Algenta Server-Sent Events surfaces — streaming simulation, agent-run event streams, and structured logs — with the exact events, frames, and how to consume them with curl and the SDKs.
- [API explorer (live)](https://docs.algenta.ai/http-api/explorer.md): Try the Algenta HTTP API live — set your key and base URL, then send real requests to every endpoint from the browser.
- [Error codes](https://docs.algenta.ai/http-api/errors.md): Every Algenta API error code in one place — generated from the codes the server emits — with what each means and the shape of the error response.
- [API stability](https://docs.algenta.ai/http-api/stability.md): Which Algenta paths carry a compatibility promise, how deprecation is announced in the response itself, and what changes without notice.
- [Execution policy & safety limits](https://docs.algenta.ai/http-api/rate-limits.md): Unlimited Algenta execution, request safety bounds, licensed worker capacity, and configurable abuse controls.
- [Connectors overview](https://docs.algenta.ai/connectors/overview.md): Connectors are configured, testable connections to your external data — databases, warehouses, object storage, REST APIs, and files — that back a governed source the engine can query.
- [Connect a database](https://docs.algenta.ai/connectors/databases.md): Connect a database directly to Algenta — the engine reads your tables under a governed contract, never copies them, and keeps credentials encrypted at rest.
- [Connectors reference](https://docs.algenta.ai/connectors/reference.md): Every supported Algenta connector type, grouped by category, with its required and optional config fields, the canonical type names, and the untested-live-error status lifecycle.
- [Snowflake](https://docs.algenta.ai/connectors/snowflake.md): Create a Snowflake connector with its real config fields, test the connection, browse its tables, onboard one as a dataset, and run a governed query against it.
- [BigQuery](https://docs.algenta.ai/connectors/bigquery.md): Connect Google BigQuery to Algenta — create a bigquery connector with project and service-account credentials, test it, browse datasets, and run a governed query.
- [ClickHouse](https://docs.algenta.ai/connectors/clickhouse.md): Connect ClickHouse to Algenta as a governed connector. Configure host, port, database, user, password, and secure, then test, browse, connect, and query.
- [PostgreSQL](https://docs.algenta.ai/connectors/postgres.md): Create a PostgreSQL connector with its real config fields, test the connection, browse its tables, onboard one as a governed dataset, and run a first governed query against it.
- [Oracle Database](https://docs.algenta.ai/connectors/oracle.md): Connect Oracle Database or Autonomous Database with Algenta's bundled Thin-mode driver, verify credentials once, browse schemas, and onboard a table or query.
- [MySQL](https://docs.algenta.ai/connectors/mysql.md): Create a MySQL connector with its real config fields, and understand the managed connection-test behavior for MySQL before you onboard data.
- [Microsoft SQL Server](https://docs.algenta.ai/connectors/mssql.md): Create a Microsoft SQL Server connector with its real config fields, and understand the managed connection-test behavior for SQL Server before you onboard data.
- [SQLite](https://docs.algenta.ai/connectors/sqlite.md): Create a SQLite connector from a database file path, test the connection, browse its tables, onboard one as a governed dataset, and run a first governed query against it.
- [Amazon Redshift](https://docs.algenta.ai/connectors/redshift.md): Create an Amazon Redshift connector with its real config fields, test the connection over the PostgreSQL wire, browse its tables, onboard one as a governed dataset, and run a first governed query.
- [Amazon S3](https://docs.algenta.ai/connectors/s3.md): Create an Amazon S3 connector with its real config fields, test object access, browse a bucket, onboard one object as a governed dataset, and run a first governed query against it.
- [Google Cloud Storage](https://docs.algenta.ai/connectors/gcs.md): Create a Google Cloud Storage connector with its real config fields, test object access, browse a bucket, onboard one object as a governed dataset, and run a first governed query.
- [Azure Blob Storage](https://docs.algenta.ai/connectors/azure-blob.md): Create an Azure Blob Storage connector with its real config fields, test blob access, browse a container, onboard one blob as a governed dataset, and run a first governed query.
- [Redis](https://docs.algenta.ai/connectors/redis.md): Create a Redis connector with its real config fields, test the connection, browse its keys, onboard selected keys as a dataset, and run a governed query against them.
- [Neo4j](https://docs.algenta.ai/connectors/neo4j.md): Create a Neo4j connector with its real config fields, test the connection, browse its labels and relationship types, onboard a selection as a dataset, and run a governed query against it.
- [Elasticsearch](https://docs.algenta.ai/connectors/elasticsearch.md): Create an Elasticsearch connector with its real config fields, test the connection, browse its indices, onboard one as a dataset, and run a governed query against it.
- [Files & uploads](https://docs.algenta.ai/connectors/files.md): Create a file connector for CSV, TSV, JSON, Excel, or Parquet data with its real config fields, test that it parses, onboard it as a dataset, and run a governed query against it.
- [Code repositories](https://docs.algenta.ai/connectors/code-repositories.md): Create a repository connector for GitHub, GitLab, Bitbucket, a local checkout, or an archive, test that the source is reachable, and browse its files to feed repository intelligence.
- [REST API](https://docs.algenta.ai/connectors/rest-api.md): Create a rest connector for any HTTP JSON API — set base\_url and auth\_type, test reachability, browse endpoints, connect a dataset, and run a governed query against it.
- [Benchmarks](https://docs.algenta.ai/benchmarks/overview.md): How Algenta measures and publishes performance — what each benchmark claims, the baseline it is measured against, and the rules every published number follows.
- [Compute kernels](https://docs.algenta.ai/benchmarks/compute-kernels.md): Quantized matrix-vector results — the fused NF4 path measured against expand-then-multiply, with correctness asserted against the reference kernel.
- [NVIDIA kernel benchmarks](https://docs.algenta.ai/benchmarks/nvidia-kernel-benchmarks.md): One portable FP32 kernel, unmodified, measured on NVIDIA T4, L4 and A100 — GEMM and saxpy with bit-exact validation; percentages are against spec-sheet peaks.
- [NVIDIA LLM throughput](https://docs.algenta.ai/benchmarks/nvidia-llm-throughput.md): Measured LLM inference throughput of the CUDA-native session on NVIDIA A100, H100, H200 and L4 — Qwen3-0.6B, Qwen2.5-7B and Gemma-3-27B prefill and decode against llama.cpp, vLLM, SGLang and TensorRT-
- [NVIDIA LLM serving](https://docs.algenta.ai/benchmarks/nvidia-llm-serving.md): Measured serving characteristics of the CUDA-native engine on NVIDIA A100, L4 and H100 — concurrent serving, single-stream decode, batched decode, prefill, cold start, VRAM, determinism; Gemma 3 27B T
- [Attention](https://docs.algenta.ai/benchmarks/attention.md): Attention kernel results — tiled attention memory against sequence length, rotary embedding cost, and the correctness checks that ran alongside every timing.
- [Normalization](https://docs.algenta.ai/benchmarks/normalization.md): Normalization kernel results — RMSNorm and LayerNorm forward and backward cost, the price of the numerically safe variance, and gradient checks against finite differences.
- [Feed-forward](https://docs.algenta.ai/benchmarks/feed-forward.md): Feed-forward and embedding kernel results — gated activation cost across three variants, the two GELU conventions and why they are not interchangeable, and embedding scatter cost.
- [Code retrieval quality](https://docs.algenta.ai/benchmarks/code-retrieval.md): Code retrieval quality on CoIR, measured with nDCG@10 against BM25 baselines evaluated on the same harness.
- [Dataframe group-by](https://docs.algenta.ai/benchmarks/dataframe-groupby.md): Group-by aggregation on the H2O db-benchmark, 0.5 GB tier, through Algenta's composed route — six of six questions ahead of both DuckDB 1.5.5 and Polars 1.43.1 on the minimum.
- [Sort](https://docs.algenta.ai/benchmarks/sort.md): LSD radix sort over Float64 keys against numpy.sort's introsort — 12 of 14 gate-clean cells ahead across three input orders, from 100,000 to 102.4 million elements.
- [Top-k selection](https://docs.algenta.ai/benchmarks/top-k.md): Bounded top-k selection over a flat Float64 column, against numpy.argpartition — a real loss at the smallest size tested, and a 5x-163x win at every size above it.
- [Graph betweenness centrality](https://docs.algenta.ai/benchmarks/graph-betweenness.md): Brandes betweenness centrality over a CSR adjacency, against networkx — ahead in every one of 6 cells, from 3.0x at 200 nodes to 4.3x at 2,000.
- [Survival analysis](https://docs.algenta.ai/benchmarks/survival-analysis.md): Kaplan-Meier and Cox proportional hazards, against lifelines 0.30.0, on two current real-world datasets plus synthetic scaling — wins from 3.0x to 142.9x, one loss disclosed.
- [Least-squares accuracy](https://docs.algenta.ai/benchmarks/least-squares-accuracy.md): Least-squares accuracy on the NIST StRD linear-regression suite, measured in correct digits against the certified values, against fourteen reference arms on the same data.
- [Forecasting on the joint axis](https://docs.algenta.ai/benchmarks/forecasting-joint-axis.md): Forecast accuracy and latency measured together on all 100,000 M4 series against statsforecast 2.1.1 — equal or better accuracy at 2.42x-3.99x the speed.
- [FinQA exact execution](https://docs.algenta.ai/benchmarks/finqa-exact-execution.md): Execution accuracy on all 8,281 gold programs in the real FinQA dataset — Algenta's exact-arithmetic driver against three naive-reimplementation baselines, on the dataset's own exact-match rule.
- [LLM inference overview](https://docs.algenta.ai/benchmarks/llm-inference-overview.md): LLM inference on Apple silicon — prefill, single-stream decode, and batched decode for Qwen3-0.6B-Base, measured against llama.cpp and mlx-lm.
- [Prefill (token input) throughput](https://docs.algenta.ai/benchmarks/llm-prefill-throughput.md): Prompt prefill (token-input) throughput for Qwen3-0.6B-Base on Apple silicon — a 256-token prompt at ~10,700 tok/s, ahead of llama.cpp's ~10,199, on any prompt length.
- [Decode throughput](https://docs.algenta.ai/benchmarks/decode-throughput.md): Decode tokens per second for Qwen3-0.6B-Base on Apple silicon — 324.8 tok/s median vs mlx-lm 281.6 and llama.cpp 254.0, interleaved order-rotated rounds on the same machine.
- [Batched decode throughput](https://docs.algenta.ai/benchmarks/batched-decode-throughput.md): Batched (continuous-batching) decode for Qwen3-0.6B-Base on Apple silicon — many sequences per weight sweep. 2.1-2.6x llama.cpp concurrent throughput at B=8-32, token-verified.
- [LLM tool-calling uplift](https://docs.algenta.ai/benchmarks/llm-tool-calling-uplift.md): gpt-5-mini answering 16 real questions against a 33,125-row CSV, with and without Algenta's tool-calling access to its own compiled engine — 16 of 16 correct with tools, 1 of 16 without.
- [Troubleshooting](https://docs.algenta.ai/help/troubleshooting.md): Hit a snag? Find your message, copy the fix, get back to work — most Algenta issues are a missing key, an engine it can't reach, or a license flag.
- [Glossary](https://docs.algenta.ai/help/glossary.md): Every Algenta term in one place — the engine, the surfaces, decision planning, governed data, agentic runs, deployment, and licensing vocabulary, each defined once.
- [Quick reference](https://docs.algenta.ai/help/quick-reference.md): One-page cheat sheet — every common Algenta task mapped to its HTTP endpoint, de CLI command, and Python SDK call, plus the handful of env vars you need most.
- [Using these docs with AI](https://docs.algenta.ai/help/ai-access.md): Use the Algenta docs with any AI — the auto-generated llms.txt context files, per-page Markdown, the built-in MCP server, and copy/open-in-chat actions.
- [Changelog](https://docs.algenta.ai/help/changelog.md): What's new in Algenta — notable releases, improvements, and fixes to the engine, SDKs, MCP server, and self-hosted deployment, newest first.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on a page URL with the `ask` query parameter:
```
GET https://docs.algenta.ai/overview.md?ask=<question>
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.