| title | Introduction |
|---|---|
| description | Knowledge Stack is a REST API for ingesting, organizing, and semantically searching documents across multi-tenant workspaces. |
Knowledge Stack gives you a complete backend for knowledge management. You ingest documents, Knowledge Stack processes and chunks them, and you query that knowledge with semantic search or AI-powered conversations — all through a clean REST API.
Upload files or raw content via `POST /v1/documents/ingest`. Knowledge Stack processes, chunks, and indexes them automatically. Ingesting a new version of an existing document is equally straightforward. Run natural-language queries against your knowledge base with `POST /v1/chunks/search`. Results are ranked by semantic relevance, not just keyword overlap. Create a thread and send user messages. Knowledge Stack retrieves relevant context from your documents and streams an AI-generated response — no RAG plumbing required on your end. Every resource belongs to a tenant. You can create and manage multiple tenants, keeping each customer's knowledge completely isolated. Control who can read or write specific parts of your knowledge base using path-level permissions. Group users together and assign permissions to the group. Define ingestion workflows and invoke them on demand. Monitor runs, inspect status, and rerun failed jobs through the Workflows API.All API requests go to:
https://api-staging.knowledgestack.ai
All endpoints are versioned under /v1. A full request path looks like:
https://api-staging.knowledgestack.ai/v1/documents/ingest
The current API version is v1. The version is part of every endpoint path, not a header.
Make your first API call in minutes — ingest a document, run a search, and start a conversation. Learn how to get an API key and authenticate your requests.