How to Manage Fast.io Intelligence Mode via API

What is Fast.io Intelligence Mode?

Fast.io Intelligence Mode automatically indexes workspace files for semantic search, accessible directly via the API.

When enabled on a workspace, uploads trigger background processing. Documents and code get text extraction, chunking, embedding, and summarization. Files reach "ready" state for RAG queries.

Key capabilities include natural language search like "Q3 contracts with Acme" and chat across folders with source citations. It works on the free agent tier alongside 50GB storage.

Competitors require Pinecone or custom DBs. Fast.io bundles it as a toggle, saving setup time.

Check Fast.io AI docs for chat endpoints.

Prerequisites for API Management

Sign up at fast.io for an agent account (no credit card). Generate API key at Settings > API Keys.

Authenticate requests with Authorization: Bearer {token}. Base URL: https://api.fast.io/current/.

Create a test workspace:

curl -X POST "https://api.fast.io/current/org/{org_id}/workspaces/" \\
  -H "Authorization: Bearer {token}" \\
  -d "folder_name=test-ws" \\
  -d "name=Test Intelligence Workspace"

Note the profile_id. Use it for updates.

Enable Intelligence Mode

Use PATCH to toggle:

curl -X PATCH "https://api.fast.io/current/workspaces/{ws_id}/" \\
  -H "Authorization: Bearer {token}" \\
  -d "intelligence=true"

Response confirms "intelligence": true. Existing files reprocess asynchronously.

JavaScript fetch example:

await fetch(`https://api.fast.io/current/workspaces/${wsId}/`, {
  method: 'PATCH',
  headers: {
    'Authorization': `Bearer ${token}`,
    'Content-Type': 'application/x-www-form-urlencoded',
  },
  body: new URLSearchParams({ intelligence: 'true' }),
});

Snippet optimized for featured results.

Verify Status and File Readiness

GET workspace details:

curl "https://api.fast.io/current/workspaces/{ws_id}/" \\
  -H "Authorization: Bearer {token}"

Check "intelligence": true. List files:

curl "https://api.fast.io/current/workspaces/{ws_id}/storage/root/?page_size=20" \\
  -H "Authorization: Bearer {token}"

Files show ai_state: "ready"|"pending"|"in_progress". Poll until ready. ai.attach: true for attachments.

Query with RAG After Enabling

POST to chat endpoint:

curl -X POST "https://api.fast.io/current/workspaces/{ws_id}/ai/chat/" \\
  -H "Authorization: Bearer {token}" \\
  -d "type=chat_with_files" \\
  -d "query_text=What are payment terms?"

Scoped to workspace by default. Use folders_scope=nodeId:3 for folders.

Response streams with citations to nodeId/page/snippet.

Disable Intelligence Mode

PATCH intelligence=false. Stops new indexing, keeps existing. Use for storage-only workspaces to save credits (10/page ingestion).

Monitor usage: GET /org/{org_id}/billing/usage/limits/credits/.

Best Practices and Troubleshooting

Enable post-creation to avoid reindexing large ws. Test with small files first.

Errors: 413 too large (1GB max), 429 rate limit (headers show remaining).

Credits: Document ingestion 10/page. Agent tier resets monthly.

Production: Use webhooks for index complete, retries with exponential backoff.