All API requests require an x-api-key header. Get your key from API Keys.

Organize your documents into folders. Each folder scopes the RAG search — queries within a folder only search documents in that folder.

Upload a document (PDF, TXT, MD, Word .doc/.docx, or CSV). The file is processed asynchronously — text is extracted, chunked, and embedded for RAG queries. Returns 202 Accepted immediately.

{
  "message": "Document uploaded and queued for processing",
  "document": {
    "id": "d1e2f3a4-b5c6-...",
    "fileName": "document.pdf",
    "status": "processing",
    "folderId": "f47ac10b-...",
    "ups": ["projeto-x", "cliente-acme"]
  },
  "jobId": "job-uuid",
  "checkStatusUrl": "/api/v1/queue/jobs/job-uuid"
}

Documents in a folder can be linked to each other, forming a knowledge graph. Links are created during upload via linkedDocumentIds or ups (auto-link groups), or manually via the graph API. The graph helps AI agents understand document relationships.

# 1. Get the graph
curl ${baseUrl}/folders/FOLDER_UUID/graph   -H "x-api-key: YOUR_API_KEY"

# 2. Read chunks of a linked document
curl ${baseUrl}/folders/FOLDER_UUID/graph/nodes/DOC_UUID/chunks   -H "x-api-key: YOUR_API_KEY"

Connect remote LLMs, AI agents, or MCP clients to query your RAG via the Model Context Protocol. Uses the same authentication (API key) and rate limits as the REST API.

# 1. Discover folders
curl -X POST ${baseUrl}/mcp \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "rag_list_folders",
      "arguments": {}
    }
  }'

# 2. Explore document relationships in a folder
curl -X POST ${baseUrl}/mcp \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/call",
    "params": {
      "name": "rag_graph_explore",
      "arguments": { "folderId": "FOLDER_UUID" }
    }
  }'

# 3. Read content of linked documents
curl -X POST ${baseUrl}/mcp \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 3,
    "method": "tools/call",
    "params": {
      "name": "rag_read_document",
      "arguments": {
        "folderId": "FOLDER_UUID",
        "documentIds": ["DOC_UUID_1", "DOC_UUID_2"]
      }
    }
  }'

# 4. Query the RAG for answers
curl -X POST ${baseUrl}/mcp \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 4,
    "method": "tools/call",
    "params": {
      "name": "rag_query",
      "arguments": { "question": "What is the contract duration?", "folderId": "FOLDER_UUID" }
    }
  }'

Ask a question and get relevant document chunks (retrieve-only, default) or an AI-generated answer (when synthesize: true). By default the endpoint only runs vector similarity search and returns sources and confidence — no LLM call.

{
  "sources": [
    { "chunkId": "...", "content": "...", "distance": 0.12, "documentId": "..." }
  ],
  "confidence": 0.88
}

For chat interfaces that need conversation history. Create a conversation once, then send multiple questions — the LLM receives the full message history for context-aware answers.

Receive real-time notifications when events happen in your account. Configure endpoints to receive HTTP POST callbacks with event payloads.

Deliveries are retried up to 5 times with exponential backoff. Each delivery includes an HMAC-SHA256 signature for verification.

Rate limit headers are included in every response: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset