Skip to content

Build with Council models.

An OpenAI compatible API for chat, streaming, and retrieval over your own text, with structured citations on every grounded answer. HIPAA compliant* under a signed BAA.

  • OpenAI compatible
  • 211 million medical vectors
  • $10 in starting credit
from openai import OpenAI

client = OpenAI(
    base_url="https://developer.council.health/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="YOUR_MODEL_ID",  # GET /v1/models lists yours
    messages=[{
        "role": "user",
        "content": "Summarize first line treatment for hypertension.",
    }],
    extra_body={
        "citations": {
            "enabled": True,
            "sources": ["public_kb"],
            "top_k": 3,
        }
    },
)

print(response.choices[0].message.content)
Sample response

Summarize first line treatment for hypertension.

Guidelines recommend lifestyle changes for everyone with high blood pressure. When medication is needed, first line options include thiazide diuretics, ACE inhibitors, ARBs, and calcium channel blockers [1][2]. The choice depends on age, other conditions, and blood pressure level [1].

  1. 1Clinical practice guidelinePublic knowledge base
  2. 2Evidence reviewPublic knowledge base

Sample response for illustration. Not medical advice.

Works with the tools you already use

  • OpenAI SDKs
  • REST and JSON
  • Server sent events
  • Scoped API keys

Your first grounded request

Three steps from an empty project to an answer with sources attached.

  1. Create a key

    Sign up in the developer console and create a key scoped to what your app needs.

    developer.council.health
  2. Point your client at Council

    Keep the OpenAI SDK you use today and change the base URL.

    base_url="https://developer.council.health/v1"
  3. Ask for citations

    Add a citations object and choose the knowledge scopes your key can read.

    "citations": { "enabled": true }

Start with chat. Add evidence when you need it.

Begin with a familiar completion API, then add retrieval and your own knowledge without stitching together a separate stack.

  • A familiar chat API

    Use OpenAI SDKs, streaming responses, and the chat completion shape your team already knows.

  • Answers with traceable evidence

    Request public, customer, or personal knowledge scopes and receive structured citation metadata with the answer.

  • Bring your own knowledge

    Upload text, track ingestion, and search an account scoped knowledge base through dedicated endpoints.

Core endpoints

  • GET/v1/modelsList the models your key can call.
  • POST/v1/chat/completionsChat completions, standard or streamed, with optional citations.
  • POST/v1/kb/documentsAdd text to a knowledge base and follow its ingestion.
  • POST/v1/kb/searchSearch a knowledge base directly.
See the API reference

Safeguards on every key

The public API is separated from the private model runtime, and authorization follows the API key, not caller supplied identity fields. The API is HIPAA compliant* for protected health information handled under a signed BAA.

  • Scoped, revocable keys

    Control chat, search, document, and billing access independently for each API key.

  • Private model runtime

    The model host stays on private infrastructure and is never exposed directly.

  • Measured operations

    Per key rate limits, usage records, spend visibility, and request level operational metadata.

  • Minimal logging

    By default, API logs do not store raw prompts, completions, uploaded text, or retrieved passages.

What happens when you ask for citations

  1. Upload

    Send text to your knowledge base.

  2. Index

    Council splits and embeds it.

  3. Retrieve

    Each request finds the relevant passages.

  4. Answer

    The model replies and cites its sources.

Put grounded answers inside your product

A foundation for clinical experiences where evidence and provenance matter.

  • Clinical copilots

    Add evidence aware assistance to clinician facing workflows without rebuilding your application stack.

  • Patient education

    Generate clear, contextual explanations grounded in sources your organization can inspect.

  • Medical knowledge search

    Retrieve relevant passages from the Council public knowledge base or from content you manage.

  • Research synthesis

    Turn a question and a set of sources into a concise starting point with citations attached.

Pay for what your product uses

Start with $10 in API credits. Use core chat when generation is enough, or add Council knowledge for answers with citations.

Core chat

OpenAI compatible completions

For generation workflows that do not need retrieval.

input per 1M tokens
$2
output per 1M tokens
$10
  • Standard chat completion shape
  • Streaming with server sent events
  • Usage and spend monitoring
Get an API key
  • Credits are enforced at the account level.
  • Purchases are managed securely through the developer console.

Before your first request

Yes. Point an OpenAI compatible client at https://developer.council.health/v1 and authenticate with a Council API key. Chat completions support both standard and streaming responses.

Add a citations object to a chat request and choose the knowledge scopes your key is allowed to use. Council retrieves relevant context and returns citation metadata with the completion. Set citations.required to true when retrieval must succeed.

Yes. Keys with document permissions can upload text, follow the ingestion job, and query customer or personal knowledge scopes. The authenticated key determines the account boundary.

Yes, once a Business Associate Agreement is signed. The API is HIPAA compliant* for protected health information handled under a signed BAA, so request one before you send any.

The developer console at developer.council.health lets you create and revoke keys, review request volume and spend, purchase credits, and copy integration examples.

Make your first request

Create an account, generate a key, and make your first Council API call.

Want the whole tool built instead? Learn more

For clinical support workflows. Qualified professionals remain responsible for care decisions.