Fyi Cli

What the edithatogo/fyi-cli MCP server does

The edithatogo/fyi-cli MCP server provides an MCP interface for managing Freedom of Information and Official Information requests on Alaveteli-based platforms. It is intended for work spanning multiple jurisdictions, including New Zealand, Australia, and the UK. The supported platform model also covers other Alaveteli deployments, such as WhatDoTheyKnow, FYI.org.nz, RightToKnow, and FragDenStaat.

The project stores application data in a local SQLite database. Its documented scope includes request records, authorities, correspondence, offline synchronization, and service health checks. The broader project can also submit and archive requests, monitor instances for updates, generate reports, and export data in JSON, CSV, and HTML formats. WARC/WACZ capture is listed for archiving requests, correspondence, and attachments.

How it works

The repository is organized as a Rust workspace containing the core library, CLI, and MCP components. The MCP component makes the request-management functionality available to compatible AI clients, while the command-line component supports direct local operation.

A typical workflow starts by initializing the database, registering a request with an authority, and then listing or managing tracked requests. The CLI quick start shows init-db, register-request, and list-requests as the initial commands. The MCP interface is suitable when an AI assistant needs to inspect or update the same locally managed request data.

The project describes privacy-oriented options including local storage, optional encryption, Tor support, encrypted credentials, and operating-system keyring integration. The supplied material does not specify which options are exposed by individual MCP tools or what configuration each option requires.

Setup and configuration

The recommended Rust installation builds the CLI from the workspace with:

cargo install --path crates/fyi-cli

The MCP binary can be installed from its workspace path with:

cargo install --path crates/fyi-mcp

After installing the CLI, initialize the local database with fyi-cli init-db. The repository also documents a crates.io installation path for fyi-cli, but notes that this is available once the package is published. A legacy Python package exists, although the README says it is a reference implementation and is no longer being extended.

The project documentation includes separate installation, configuration, API-key setup, troubleshooting, and user-guide documents. The provided material does not identify a required environment variable or a specific credential needed for the MCP server, so deployment-specific configuration should be checked in those documents.

Tools and capabilities

The edithatogo/fyi-cli MCP server is described as providing tools for:

• Creating and tracking FOI/OIA requests
• Managing authorities and correspondence
• Synchronizing data while offline
• Checking platform health
• Monitoring Alaveteli instances for updates
• Supporting request and correspondence archiving

The project also lists reporting, analytics, handover documents, exports, and encrypted storage among its broader capabilities. Exact MCP tool names and parameter schemas are not included in the supplied excerpt.

Limitations and notes

The service is designed for Alaveteli-based platforms rather than a general-purpose records system. Platform behavior may vary by jurisdiction and deployment. The README describes localization as forthcoming, so jurisdiction-aware metadata is available in the project’s stated scope but further localization work is planned.

The official MCP Registry entry is io.github.edithatogo/fyi-mcp, listed at version 0.1.2 in the repository excerpt. Claude connector assets are described as prepared, but hosted HTTPS deployment and external submission were still pending. Container assets and several package-manager routes are likewise described as ready or planned rather than universally live. The edithatogo/fyi-cli MCP server should therefore be installed using a confirmed local Rust or release path for the target environment.