Private AI Document Search on AWS

Built for your workflow

Teams managing runbooks, technical documentation or internal knowledge files can retrieve relevant passages through a browser interface. Private retrieval mode works without an external LLM key; generated answers require an optional model-provider configuration.

What’s included

Document upload and conversion for PDFs, Word documents, text and Markdown.
Qdrant vector indexing and semantic retrieval with source names and similarity scores.
Persistent indexed document library and browser search interface.
First-boot credentials, Basic Auth protection, status helpers and backup tooling.

What Code Creator adds

Code Creator integrates the document conversion, vector storage and search interface into a Docker Compose environment. First-boot provisioning and operating tools provide a starting point for document retrieval without assembling those services separately.

From launch to first use

Launch according to the listing and retrieve the generated access information over SSH.
Allow three to five minutes for Docling initialization; the web page may appear before conversion is ready.
Open the protected interface, upload a representative document and test passage retrieval.

Use the current Marketplace usage instructions for the full launch procedure and required instance settings.

Before you launch

The initial deployment uses HTTP with Basic Auth; configure HTTPS and restrict access before sending sensitive material. Optional external model providers change the data flow and may add charges. The listing describes MCP-ready positioning, so this draft does not promise a completed agent integration.

Review current software pricing, AWS infrastructure costs and the supported configuration on AWS Marketplace.