Skip to content
MCP server

Laserfiche MCP

Search and read self-hosted Laserfiche documents from Claude, read-only by default

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI, self-hosted, Windows
Runs
Locally
Version
2.6.2
Listing updated
today

Similar MCP servers

About Laserfiche MCP

Laserfiche MCP is an open-source MCP server for files and documents. It runs locally on your machine and is started by your AI client.

This server connects Claude Desktop, Claude Code or any MCP client to a self-hosted Laserfiche repository, and the same program also works as a command-line tool with commands such as ls, cat, search and dedupe. It is owned by the General Conference of Seventh-day Adventists and not affiliated with Laserfiche.

Search uses native Laserfiche queries or name patterns, can draft queries from your own templates and fields, and returns matching passages with page numbers from the OCR index, so many questions about scans need no download. Claude can also view images such as PNG and JPEG to describe and tag them, with a token-cost warning first. Text extraction covers PDF, Word, PowerPoint, email, HTML, RTF and plain text, and the server can compare entries and find byte-identical duplicates.

Writes are off until LF_READ_ONLY is set to false. When enabled, it can create, copy, move, rename, edit fields and tags and delete, with previews for deletes, renames and moves, folder fencing, allow-listed tools and capped bulk deletes. Laserfiche Cloud support is beta and unverified against a live tenant. It installs from PyPI or as a Desktop extension.

Install

Run in a terminal

claude mcp add laserfiche-mcp --env LF_REPO_API_URL='<LF_REPO_API_URL>' --env LF_REPOSITORY_ID='<LF_REPOSITORY_ID>' --env LF_USERNAME='<LF_USERNAME>' --env LF_PASSWORD='<LF_PASSWORD>' -- uvx laserfiche-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "laserfiche-mcp": {
      "command": "uvx",
      "args": [
        "laserfiche-mcp"
      ],
      "env": {
        "LF_REPO_API_URL": "<LF_REPO_API_URL>",
        "LF_REPOSITORY_ID": "<LF_REPOSITORY_ID>",
        "LF_USERNAME": "<LF_USERNAME>",
        "LF_PASSWORD": "<LF_PASSWORD>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "laserfiche-mcp": {
      "command": "uvx",
      "args": [
        "laserfiche-mcp"
      ],
      "env": {
        "LF_REPO_API_URL": "<LF_REPO_API_URL>",
        "LF_REPOSITORY_ID": "<LF_REPOSITORY_ID>",
        "LF_USERNAME": "<LF_USERNAME>",
        "LF_PASSWORD": "<LF_PASSWORD>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "laserfiche-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "laserfiche-mcp"
      ],
      "env": {
        "LF_REPO_API_URL": "<LF_REPO_API_URL>",
        "LF_REPOSITORY_ID": "<LF_REPOSITORY_ID>",
        "LF_USERNAME": "<LF_USERNAME>",
        "LF_PASSWORD": "<LF_PASSWORD>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "laserfiche-mcp": {
      "command": "uvx",
      "args": [
        "laserfiche-mcp"
      ],
      "env": {
        "LF_REPO_API_URL": "<LF_REPO_API_URL>",
        "LF_REPOSITORY_ID": "<LF_REPOSITORY_ID>",
        "LF_USERNAME": "<LF_USERNAME>",
        "LF_PASSWORD": "<LF_PASSWORD>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "laserfiche-mcp": {
      "command": "uvx",
      "args": [
        "laserfiche-mcp"
      ],
      "env": {
        "LF_REPO_API_URL": "<LF_REPO_API_URL>",
        "LF_REPOSITORY_ID": "<LF_REPOSITORY_ID>",
        "LF_USERNAME": "<LF_USERNAME>",
        "LF_PASSWORD": "<LF_PASSWORD>"
      }
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Configuration

  • LF_REPO_API_URLBase URL of your Laserfiche Repository API Server, e.g. https://lf.example.com/LFRepositoryAPIRequired
  • LF_REPOSITORY_IDRepository name or ID to operate against.Required
  • LF_USERNAMEService-account username with read access to the repository.Required
  • LF_PASSWORDService-account password.Required
  • LF_API_VERSIONRepository API version: v1 (default) or v2.
  • LF_READ_ONLYRead-only mode. Defaults to true; set false to register the write tools.

What it does

OCR passage search
Returns matching passages with page numbers from the OCR index.
Document reading
Extracts text from PDF, Office, email and other formats.
Image viewing
Lets Claude describe and tag images, after a cost warning.
Duplicate finding
Detects byte-identical files by size then content.
Opt-in writes
Creates, edits, moves and deletes only when enabled.
Safety guardrails
Previews destructive actions and caps bulk deletes.
Command-line mode
Runs the same operations without AI.

Good for

  • Find which leases mention an unpaid balance
  • Summarise the newest document in a folder
  • Compare two scans to see whether they are the same record
  • Find duplicate invoices and the space they use
  • Approve field values across several entries

Pricing

Open source

Free, open source

Laserfiche MCP FAQ

Is it read-only?

By default yes. Writes require LF_READ_ONLY=false and destructive actions need a second confirmation.

Does it support Laserfiche Cloud?

Support is beta and has never been verified against a live tenant.

What does it need?

Python 3.10 or newer, a reachable Repository API Server and an account that can read the repository.

Can I use it without Claude?

Yes. The same operations are available as a command-line tool with JSON output.

Is Laserfiche MCP free?

Yes. Laserfiche MCP is open source, so you can use and self-host it for free.

How do I add Laserfiche MCP to Claude Code?

Run this command in your terminal: claude mcp add laserfiche-mcp --env LF_REPO_API_URL='<LF_REPO_API_URL>' --env LF_REPOSITORY_ID='<LF_REPOSITORY_ID>' --env LF_USERNAME='<LF_USERNAME>' --env LF_PASSWORD='<LF_PASSWORD>' -- uvx laserfiche-mcp

What are the best alternatives to Laserfiche MCP?

The closest alternatives to Laserfiche MCP are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and NeuroStack.

Where this comes from

  • Description: written from the project's README.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers