Skip to main content

Ask API overview

End user & admin

Everything you do in the Ask web app (chat with agents, upload documents, manage collections) you can also do over a REST API. This page tells you what the API covers and where the full reference lives; it does not duplicate the reference itself.

The API reference is the source of truth

Request/response schemas, parameters, and code samples live in the API Reference. This guide only covers what the app UI adds around it, mainly API keys, and gives you enough orientation to know which reference page to open.

What you can build​

The API exposes the same four building blocks you see in the app, so a custom integration can go anywhere the web app can't:

  • Chatbots and virtual agents: embed an Ask agent in your own product or internal tool.
  • Document Q&A: push documents in programmatically and query them through an agent.
  • Automation: script bulk work: create agents, sync collections, pull conversation history into other systems.
  • Custom front ends: build your own UI on top of Ask's orchestration, with streaming responses.

What the API covers​

All endpoints are versioned under /v1/. The groups mirror the app's main surfaces:

GroupWhat it doesReference
AgentsList, create, update, and delete agents; fetch an agent's ID and configuration.Agents endpoints
ConversationsSend a message to an agent (streaming or not), continue a conversation, send messages with file uploads, list history and messages.Conversations endpoints
CollectionsCreate and manage knowledge collections, share them with users or groups.Collections endpoints
DocumentsUpload documents into collections, track ingestion, retry or cancel processing, fetch download URLs.Documents endpoints
Users & groupsSearch users and groups in your org, used when sharing collections programmatically.Search endpoints

Authentication in one paragraph​

Every request is authenticated with an API key you create yourself in Personal settings → API keys, passed in the x-api-key header. Keys act on your behalf, capped by your own permissions, and can be scoped down per key. See Create & manage API keys for the full walkthrough.

bash
curl -X GET "https://api.ask.dev.ai71services.ai/v1/agents/" \
-H "x-api-key: $ASK_API_KEY"
Base URL

Every example on this page targets https://api.ask.dev.ai71services.ai, the API host for the deployment you are reading these docs on. Each environment has its own, and these docs are published alongside the deployment they describe, so the URLs here are always the right ones for the instance you signed in to.

Your first call​

Don't start from this page. The Getting Started guide walks you through the whole flow in five steps, with Python and Node.js examples:

  1. Generate an API key (also covered in this guide's key page).

  2. Create an agent in the app and note its Agent ID, or fetch it with GET /v1/agents/.

  3. Send a message with POST /v1/conversations/.

  4. Set "stream": true for incremental, real-time responses.

  5. Pass the returned conversation_id back in to continue the same conversation.

OpenAPI spec & tooling​

The complete machine-readable spec is published at https://api.ask.dev.ai71services.ai/openapi.json (OpenAPI 3.0). Import it into Postman or Swagger UI, or generate a typed client SDK with OpenAPI Generator. The OpenAPI Specification page shows the exact commands.