Connectors & data

TypeScript SDK methods for connectors, datasets, source registration, and repository intelligence — create, test, browse, onboard, snapshot, triage, and apply.

These AlgentaClient methods bring data into the engine and turn it into governed evidence: connectors (create, test, browse, update, delete), datasets (connect, list, summarize, refresh, delete), source registration, and repository intelligence (snapshot, triage, graph, simulate, apply). Method names are camelCase; bodies stay snake_case.

Connector methods

Create a connector, test it against the live source, then browse the tables it exposes. The preview* variants test or browse an unsaved connector from an inline config, without persisting it first.

import { AlgentaClient } from "algenta-sdk";

const client = new AlgentaClient({ apiKey: process.env.ALGENTA_API_KEY });

// 1. Create a connector.
const connector = await client.createConnector({
  name: "Analytics Warehouse",
  connector_type: "snowflake",
  config: {
    account: "xy12345.eu-central-1",
    user: "ANALYTICS",
    password: process.env.SNOWFLAKE_PASSWORD,
    warehouse: "COMPUTE_WH",
    database: "SALES",
    schema: "PUBLIC",
  },
});

// 2. Test the live connection, then browse its tables.
const test = await client.testConnector(connector.id);
if (test.success) {
  const browse = await client.browseConnector(connector.id);
  console.log(browse.total, "tables");
}

Method	HTTP	Notes
createConnector(request)	POST /v1/connectors	name, connector_type, and a config; returns ConnectorInfo with status: "untested".
listConnectors(options?)	GET /v1/connectors	Paginate with page, limit.
getConnector(connectorId)	GET /v1/connectors/{id}	Fetch one connector.
updateConnector(connectorId, request)	PATCH /v1/connectors/{id}	Update name, description, visibility, or config.
testConnector(connectorId)	POST /v1/connectors/{id}/test	Real connection health check; flips status to live or error.
previewTestConnector(request)	POST /v1/connectors/test	Test an unsaved connector from a { connector_type, config } body.
browseConnector(connectorId)	GET /v1/connectors/{id}/browse	List the tables a live connector exposes.
previewBrowseConnector(request)	POST /v1/connectors/browse	Browse an unsaved connector from a { connector_type, config } body.
deleteConnector(connectorId)	DELETE /v1/connectors/{id}	Remove a connector; resolves to void.

browseConnector requires a live connector. Call testConnector first — browsing a connector that has not passed its test returns a not-connected error.

Dataset methods

connectData onboards a selection (a table or a SQL query) from a saved connector as a queryable dataset. The rest of the family lists, inspects, refreshes, and deletes datasets. listDatasets supports search, status, sourceName (sent as source_name), and compact.

Method	HTTP	Notes
connectData(request)	POST /v1/data/connect	Onboard a selection from a connection_id; returns a DatasetConnectResult.
listDatasets(options?)	GET /v1/data	Filter with search, status, sourceName; compact returns lighter rows.
getDataset(datasetId)	GET /v1/data/{id}	Full dataset detail.
getDatasetSummary(datasetId)	GET /v1/data/{id}/summary	Schema, fields, and row counts.
refreshDataset(datasetId)	POST /v1/data/{id}/refresh	Re-ingest the dataset; returns a DatasetConnectResult.
deleteDataset(datasetId)	DELETE /v1/data/{id}	Remove a dataset.

// Onboard a table from a saved connector.
const dataset = await client.connectData({
  connection_type: "database",
  provider: "snowflake",
  dataset_name: "sales_orders",
  connection_id: connector.id,
  selection: { table: "PUBLIC.ORDERS" },
  visibility: "private",
});

// Inspect what was onboarded.
const summary = await client.getDatasetSummary(dataset.dataset_id);
console.log(summary.dataset_id, summary.name);

Source registration methods

registerSource registers a raw source object as a dataset in one call; the optional second argument sets a description and name. refreshSource re-ingests a registered source by its dataset id.

Method	HTTP	Notes
registerSource(source, options?)	POST /v1/sources/register	source is the source object; options sets { description, name }.
refreshSource(datasetId)	POST /v1/data/{id}/refresh	Re-ingest a registered source; returns a SourceRegistrationResponse.

Repository intelligence methods

Repository intelligence turns a code repository into governed evidence. Snapshot the repo, triage an issue into ranked evidence, walk the symbol graph, simulate a candidate decision plan, and apply the result as a patch, branch, or pull request. Start by checking which languages the engine supports.

const capabilities = await client.getRepositoryIntelligenceCapabilities();
console.log(capabilities.supported_languages.length);

// 1. Snapshot the repository.
const snapshot = await client.createRepositorySnapshot("repo_42", {
  ref: "main",
  include_patterns: ["src/**"],
  max_files: 5000,
});

// 2. Triage an issue into ranked evidence.
const triage = await client.triageRepository("repo_42", {
  snapshot_id: snapshot.snapshot_id,
  signals: {
    issue_text: "Null pointer when the cart is empty at checkout",
    failing_tests: ["test_checkout_empty_cart"],
  },
  max_evidence_items: 20,
});

// 3. Walk the symbol graph around a symbol.
const graph = await client.queryRepositoryGraph("repo_42", {
  snapshot_id: snapshot.snapshot_id,
  symbol_name: "checkout",
  direction: "both",
  max_depth: 2,
});
console.log(triage.evidence?.length, graph.nodes?.length);

Method	HTTP	Notes
getRepositoryIntelligenceCapabilities()	GET /v1/repositories/capabilities	Supported languages and support progress.
createRepositorySnapshot(repositoryId, request)	POST /v1/repositories/{id}/snapshots	ref and optional include_patterns, exclude_patterns, max_files.
getRepositorySnapshot(repositoryId, snapshotId)	GET /v1/repositories/{id}/snapshots/{snapshotId}	Fetch one snapshot.
triageRepository(repositoryId, request)	POST /v1/repositories/{id}/triage	Ranked evidence from signals (issue_text, failing_tests, changed_files).
createRepositoryDecisionPlan(repositoryId, request)	POST /v1/repositories/{id}/decision-plans	Plan from a workspace_evidence_bundle_ref; optional model.
queryRepositoryGraph(repositoryId, request)	POST /v1/repositories/{id}/graph-query	Symbol/file graph; direction is inbound, outbound, or both.
simulateRepository(repositoryId, request)	POST /v1/repositories/{id}/simulate	Returns a DecisionEnvelope for a decision_plan_id.
applyRepository(repositoryId, request)	POST /v1/repositories/{id}/apply	mode is patch_only, local_branch, or remote_pr.

applyRepository with mode: "local_branch" or "remote_pr" writes to your repository. It requires write_permission: true and a simulation_id from a prior simulateRepository call. Use mode: "patch_only" to get the diff without writing anything.

Next steps

Connect Snowflake step by step

Create a connector and onboard a table as a dataset.

Query governed data

Run deterministic queries over onboarded datasets.

Fix an issue with repository intelligence

Snapshot, triage, simulate, and apply a fix.