Sources API

Manage data sources and RAG search.

Endpoints

Method	Endpoint	Description
GET	/api/agents/{id}/sources	List sources
POST	/api/agents/{id}/sources	Create source
POST	/api/agents/{id}/sources/crawl	Crawl a website
GET	/api/agents/{id}/sources/search	Search sources
GET	/api/agents/{id}/sources/{sourceId}	Get source
DELETE	/api/agents/{id}/sources/{sourceId}	Delete source
POST	/api/agents/{id}/sources/{sourceId}/reindex	Reindex source
PATCH	/api/agents/{id}/sources/{sourceId}	Set visibility (agent | business | contact); re-stamps chunks and reindexes — see Data Sources

List Sources

GET /api/agents/{id}/sources

Response

{
  "sources": [
    {
      "id": "source-123",
      "type": "file",
      "name": "Product Manual",
      "status": "ready",
      "chunks": 150,
      "created_at": "2024-12-01T00:00:00Z"
    },
    {
      "id": "source-456",
      "type": "crawl",
      "name": "Help Center",
      "status": "ready",
      "chunks": 320,
      "config": {
        "url": "https://help.example.com",
        "maxPages": 50,
        "maxDepth": 3,
        "pageCount": 42,
        "lastCrawledAt": "2026-03-10T14:30:00Z"
      },
      "created_at": "2026-03-01T00:00:00Z"
    }
  ]
}

Create Source

File Upload

POST /api/agents/{id}/sources
Content-Type: multipart/form-data

-F "type=file"
-F "name=Product Manual"
-F "file=@/path/to/manual.pdf"

Response

{
  "id": "source-123",
  "type": "file",
  "name": "Product Manual",
  "status": "processing",
  "created_at": "2024-12-15T10:00:00Z"
}

Crawl Website

POST /api/agents/{id}/sources/crawl
Content-Type: application/json

Starts a BFS crawl of the target website using Cloudflare Browser Rendering (/markdown + /links endpoints). The crawl runs in the background via ctx.waitUntil and does not block the response.

Request Body

{
  "url": "https://docs.example.com",
  "name": "Documentation Site",
  "maxPages": 30,
  "maxDepth": 3,
  "includePatterns": ["/docs/*", "/guides/*"],
  "excludePatterns": ["/blog/*", "/changelog/*"]
}

Parameters

Field	Type	Required	Default	Description
url	string	Yes	—	Starting URL for the crawl
name	string	Yes	—	Display name for the source
maxPages	number	No	50	Maximum pages to crawl (up to 50)
maxDepth	number	No	3	Maximum link depth from start URL (up to 5)
includePatterns	string[]	No	[]	Glob patterns — only crawl URLs matching at least one pattern
excludePatterns	string[]	No	[]	Glob patterns — skip URLs matching any pattern

Response

{
  "id": "source-789",
  "type": "crawl",
  "name": "Documentation Site",
  "status": "crawling",
  "config": {
    "url": "https://docs.example.com",
    "maxPages": 30,
    "maxDepth": 3,
    "includePatterns": ["/docs/*", "/guides/*"],
    "excludePatterns": ["/blog/*", "/changelog/*"]
  },
  "created_at": "2026-03-11T10:00:00Z"
}

The source status will be crawling while the BFS traversal is in progress. Once complete, pages are chunked and embedded, and the status transitions to ready.

Crawl Source Config Fields

When retrieving a crawl source, the config object contains these additional metadata fields:

Field	Type	Description
url	string	The starting URL
maxPages	number	Configured page limit
maxDepth	number	Configured depth limit
includePatterns	string[]	URL include patterns
excludePatterns	string[]	URL exclude patterns
pageCount	number	Total pages discovered and indexed
lastCrawledAt	string	ISO 8601 timestamp of the most recent crawl
pageHashes	object	Map of URL to SHA-256 content hash (used for delta detection)
changedPages	number	Pages that were new or modified on the last re-crawl
unchangedPages	number	Pages skipped (content hash matched previous crawl)

Reindex Source

POST /api/agents/{id}/sources/{sourceId}/reindex

Reprocesses the source content. For crawl sources, this triggers a full re-crawl with delta detection: the crawler revisits all pages but only re-indexes those whose SHA-256 content hash has changed since the last crawl. Unchanged pages are skipped.

Response

{
  "id": "source-789",
  "type": "crawl",
  "status": "crawling",
  "message": "Re-crawl started with delta detection"
}

After the re-crawl completes, the source config will include updated changedPages and unchangedPages counts:

{
  "id": "source-789",
  "type": "crawl",
  "name": "Documentation Site",
  "status": "ready",
  "config": {
    "url": "https://docs.example.com",
    "maxPages": 30,
    "maxDepth": 3,
    "pageCount": 42,
    "lastCrawledAt": "2026-03-11T15:00:00Z",
    "changedPages": 3,
    "unchangedPages": 39
  }
}

Search Sources

GET /api/agents/{id}/sources/search?query=how+to+reset+password&limit=5

Query Parameters

Parameter	Type	Description
query	string	Search query (required)
limit	number	Max results (default: 10)
source_id	string	Filter by source
min_score	number	Minimum similarity score

Response

{
  "results": [
    {
      "id": "chunk-123",
      "source_id": "source-456",
      "content": "To reset your password, go to Settings > Account > Password...",
      "score": 0.92,
      "metadata": {
        "source_name": "User Guide",
        "page": 15,
        "section": "Account Settings"
      }
    }
  ]
}

Get Source

GET /api/agents/{id}/sources/{sourceId}

Response

{
  "id": "source-123",
  "type": "file",
  "name": "Product Manual",
  "file_path": "uploads/manual.pdf",
  "status": "ready",
  "chunks": 150,
  "config": {
    "chunk_size": 1000,
    "overlap": 200
  },
  "created_at": "2024-12-01T00:00:00Z",
  "updated_at": "2024-12-15T10:00:00Z"
}

Crawl Source Response

{
  "id": "source-789",
  "type": "crawl",
  "name": "Documentation Site",
  "status": "ready",
  "chunks": 320,
  "config": {
    "url": "https://docs.example.com",
    "maxPages": 30,
    "maxDepth": 3,
    "includePatterns": ["/docs/*"],
    "excludePatterns": ["/blog/*"],
    "pageCount": 42,
    "lastCrawledAt": "2026-03-11T15:00:00Z",
    "changedPages": 3,
    "unchangedPages": 39
  },
  "created_at": "2026-03-01T00:00:00Z",
  "updated_at": "2026-03-11T15:00:00Z"
}

Delete Source

DELETE /api/agents/{id}/sources/{sourceId}

Response

{
  "success": true
}

Refresh Source

POST /api/agents/{id}/sources/{sourceId}/refresh

Reprocesses the source content. For crawl sources, use the Reindex endpoint instead, which performs a re-crawl with delta detection.

Source Types

Type	Description
file	Uploaded file
crawl	Crawled website (BFS traversal with delta detection)
database	Database query
api	API endpoint

Source Status

Status	Description
pending	Waiting to process
processing	Currently processing
crawling	Website crawl in progress (BFS traversal)
ready	Available for search
error	Processing failed
updating	Refreshing content

Supported Formats

Format	Extensions
PDF	.pdf
Text	.txt
Markdown	.md
Word	.docx
JSON	.json
CSV	.csv

Chunking Config

{
  "config": {
    "chunk_size": 1000,
    "overlap": 200,
    "method": "semantic"
  }
}

Errors

Code	Description
400	Invalid source data
404	Source not found
413	File too large
415	Unsupported format

Examples

Upload PDF

curl -X POST .../sources \
  -F "type=file" \
  -F "name=User Guide" \
  -F "file=@guide.pdf"

Search with Filter

curl ".../sources/search?query=installation&source_id=source-123&min_score=0.8"

Crawl Website

curl -X POST .../sources/crawl \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://help.example.com",
    "name": "Help Center",
    "maxPages": 30,
    "maxDepth": 3,
    "includePatterns": ["/docs/*"],
    "excludePatterns": ["/blog/*"]
  }'

Check Crawl Source Details

curl .../sources/source-789

Response includes crawl metadata (pageCount, lastCrawledAt, changedPages, unchangedPages).

Reindex Crawl Source (Delta Re-crawl)

curl -X POST .../sources/source-789/reindex

Triggers a re-crawl that only re-indexes pages whose content has changed.