Skip to main content
The AutoSage API lets you provision tenants, ingest documents, build and run agents, and stream grounded conversations directly from your own application. If you are configuring AutoSage from the dashboard instead, see the Product docs.

Base URL

All endpoints are served under a versioned prefix:
https://api.autosage.ai/api/v1
Authentication endpoints live under /api/auth, and invitation flows are served at the root. Each endpoint page notes its exact path.

Conventions

JSON everywhere

Requests and responses are JSON. Send Content-Type: application/json on requests with a body.

Bearer auth

Authenticate with an API key: Authorization: Bearer sk_.... See Authentication.

Schema-defined fields

Field naming follows each endpoint schema and can mix styles. Use the endpoint reference or OpenAPI explorer for the exact request and response shape.

Predictable errors

Errors return the right HTTP status with a JSON message. 400 validation, 401/403 auth, 404 not found, 409 conflict.

The resource hierarchy

The API mirrors the platform hierarchy. Almost every request resolves against it, and access is checked at each level:
Organization  →  Environment  →  Tenant  →  Knowledge base  →  Documents / Agents / Chats
An API key is bound to a single environment and carries the scopes and tenant/knowledge-base access it is allowed to use. Understanding this model is the key to designing a secure integration — start with Tenancy and Access & scopes.

Two ways in

Dashboard users

People sign in with email/password or Google. Session-based auth powers the dashboard and admin operations like managing members and environments.

API keys

Your application authenticates with sk_... keys to provision resources and run chats — including on behalf of your own end users.

Interactive reference

An OpenAPI-powered explorer is available at /swagger on the API host. These developer docs add the concepts, lifecycles, and guidance that the raw schema can’t express.

Essential guides

These guides begin with the product decision, then show the API implementation:

KB prompts vs. agents

Decide where shared behavior belongs and when you need a saved agent.

Chat vs. agent runs

Understand chats, agents, runs, and the chat created for each run.

Share context between KBs

Compose isolated knowledge bases through controlled delegation.

Agent in two minutes

Create version 1 and run it with the shortest valid API flow.

Next steps

1

Get a key

Create an API key in the dashboard and learn how authentication works.
2

Make your first call

Follow the quickstart to run a grounded chat end to end.
3

Go deeper

Read the lifecycles to understand agents, chat, ingestion, and memory in detail.