Archive Agent MCP Server

What is Archive Agent?

Archive Agent is an intelligent file indexer that brings RAG (Retrieval Augmented Generation) to your command line. It indexes local documents (PDFs, images, Markdown, plaintext, and more) using automatic OCR, semantic chunking, and a local Qdrant vector database, then lets you search or query those files with natural language. It integrates with AI tools via a built-in MCP server — it is not a chatbot.

How to use Archive Agent?

First, install Docker and Python ≥3.10, then run git clone, cd Archive-Agent, chmod +x install.sh, ./install.sh. The script sets up a Python environment, installs dependencies, and deploys a Qdrant Docker container. After installation, configure an AI provider (OpenAI, OpenRouter, Ollama, or LM Studio) and run archive-agent to use the CLI (track, commit, search, query, GUI) or start the MCP server (archive-agent mcp).

Key features of Archive Agent

• Semantic natural‑language search & query (RAG) on local files
• Automatic OCR for images and PDFs (experimental)
• Local indexing with self‑hosted Qdrant vector database
• Supports OpenAI, OpenRouter (400+ models), Ollama, and LM Studio
• Built‑in MCP server for workflow integration
• Fully resumable parallel processing with AI cache & retry logic
• Smart semantic chunking with context headers and reranking

Use cases of Archive Agent

• Search a personal document archive using natural‑language questions
• Query technical manuals, PDFs, or scanned images without manual browsing
• Integrate document Q&A into an AI assistant or IDE via MCP
• Index and search offline documents using local LLMs (Ollama/LM Studio)
• Automate OCR and entity extraction from scanned images on a schedule