> For the complete documentation index, see [llms.txt](https://docs.algenta.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.algenta.ai/sdks/python/llm.md).

# LLM & embeddings

These methods call Algenta's governed LLM surface: model discovery, tokenization, chat completions, the responses API, embeddings, similarity, and reranking. They are available on both `AlgentaClient` and `AsyncAlgentaClient` (await the async calls). Every model argument routes through the instance's model catalog — call `list_models()` first to see what is executable, as described in [API overview](/http-api/overview.md).

## Method reference

| Method                                                                                    | Purpose                                                                                                                          |
| ----------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| `list_models()`                                                                           | List the LLM utility models this instance can execute (`LLMModelListResult`). Reads `/v1/models`.                                |
| `tokenize(text, *, model="text.tokenizer")`                                               | Tokenize text (`TokenizeResult`). Posts to `/v1/tokenize`.                                                                       |
| `count_tokens(text, *, model="text.tokenizer")`                                           | Count tokens in text (`CountTokensResult`). Posts to `/v1/count_tokens`.                                                         |
| `chat_completions(messages, *, model="text.tokenizer")`                                   | Run a chat completion (`ChatCompletionsResult`). Posts to `/v1/chat/completions`.                                                |
| `stream_chat_completions(messages, *, model="text.tokenizer")`                            | Stream a chat completion as chunks. Posts to `/v1/chat/completions`.                                                             |
| `responses(input_value, *, model="text.tokenizer", dimensions=64)`                        | Run the responses API and return output items (`ResponsesResult`). Posts to `/v1/responses`.                                     |
| `stream_responses(input_value, *, model="text.tokenizer", dimensions=64)`                 | Stream responses events. Posts to `/v1/responses`.                                                                               |
| `embeddings(input_value, *, model="text.hash_embedding_v1", dimensions=64)`               | Compute embeddings (`EmbeddingsResult`). Posts to `/v1/embeddings`.                                                              |
| `embedding_similarity(left, right, *, model="embeddings.cosine_similarity")`              | Compute similarity between two vectors (`EmbeddingSimilarityResult`). Posts to `/v1/embeddings/similarity`.                      |
| `rerank(query_embedding, documents, *, model="embeddings.cosine_similarity", top_n=None)` | Rerank documents against a query embedding (`RerankResponseResult`). Posts to `/v1/rerank`.                                      |
| `resolve_artifact_bridge(*, repo_id, filename, revision=None, local_files_only=True)`     | Resolve a model artifact through the governed artifact bridge (`ArtifactBridgeResolveResult`). Posts to `/v1/artifacts/resolve`. |

## Example: discover a model and run a chat completion

```python
import os
from decision_engine import AlgentaClient

client = AlgentaClient(api_key=os.environ["ALGENTA_API_KEY"])

# Only models listed here are executable on this instance.
models = client.list_models()
print([m.id for m in models.data])

count = client.count_tokens("How many tokens is this?")
print(count.token_count)

reply = client.chat_completions(
    [
        {"role": "system", "content": "You are concise."},
        {"role": "user", "content": "Summarize the incident in one sentence."},
    ],
)
print(reply.choices[0].message.content)
```

## Example: embed, compare, and rerank

```python
import os
from decision_engine import AlgentaClient

client = AlgentaClient(api_key=os.environ["ALGENTA_API_KEY"])

vectors = client.embeddings(["orders pipeline", "checkout latency"])
left = vectors.data[0].embedding
right = vectors.data[1].embedding

similarity = client.embedding_similarity(left, right)
print(similarity.score)

ranked = client.rerank(
    query_embedding=left,
    documents=[
        {"id": "doc-1", "text": "checkout latency runbook"},
        {"id": "doc-2", "text": "billing reconciliation notes"},
    ],
    top_n=1,
)
for item in ranked.data:
    print(item.rank, item.id, item.score)
```

## Related pages

{% content-ref url="/pages/7stbYFWr2YSgMNff3QuV" %}
[API overview](/http-api/overview.md)
{% endcontent-ref %}

{% content-ref url="/pages/l0mqGrhl5ic9OfaGy52Q" %}
[Simulations & queries](/sdks/python/simulations-and-queries.md)
{% endcontent-ref %}

{% content-ref url="/pages/9l598Hwbo4aGEvKi96R8" %}
[MCP server](/sdks/mcp.md)
{% endcontent-ref %}


---

# 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 that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.algenta.ai/sdks/python/llm.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

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.