Importer Quick Start

Install the Importer and turn your first document into a knowledge graph stored in ArangoDB

Prerequisites

An Arango Contextual Data Platform deployment.
A project in your target database. The project name prefixes the collection names, so it must follow ArangoDB naming rules. See Projects.
LLM and embedding API access (OpenAI-compatible or Triton-compatible).
A valid JWT (Authorization: Bearer ...).
A document to import (.txt, .md, .pdf, or a supported Office file such as .docx). See Format support.

Import your first document

You reach the Importer through the platform API gateway on port 8529. Every call includes an Authorization: Bearer <token> header.

Install the Importer service

Install and start the service through the Arango Control Plane. Configure your LLM and embedding providers at install time.

POST https://<EXTERNAL_ENDPOINT>:8529/_platform/acp/v1/graphragimporter

Note the serviceIdPostfix from the response - you need it for the following calls.

Submit an import

Submit a single file (use /import-multiple for a batch). Keep the default rag_mode: "full_graphrag" to build a complete knowledge graph.

POST https://<EXTERNAL_ENDPOINT>:8529/graphrag/importer/:serviceIdPostfix/v1/import

Monitor progress

For single-file imports, watch the platform service status. For multi-file imports, poll the job:

GET https://<EXTERNAL_ENDPOINT>:8529/graphrag/importer/:serviceIdPostfix/v1/jobs/:job_id

Verify the result

Open your ArangoDB database and confirm the knowledge-graph collections (prefixed with your project name) were created and populated. See Verify and explore.

You now have a knowledge graph in ArangoDB. Query it with AutoRAG or with AQL directly.

Next steps

LLM Configuration: Configure chat and embedding providers.
Import Files: Single-file and multi-file workflows.
Architecture: Collections, vector indexes, and the async-job lifecycle.