maha-strategies/maha-mcp-bridge — explained in plain English
Analysis updated 2026-05-18
Let an AI coding agent query a local database without exposing raw data.
Redact personal information from text before it reaches an external API.
Run compliance audits on passages of text using a hash chained ledger.
Mount a purchased book so an AI agent can search it section by section.
| maha-strategies/maha-mcp-bridge | 0xkinno/neuralvault | 0xmayurrr/ai-contractauditor | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | hard | easy |
| Complexity | 3/5 | 4/5 | 2/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a Maha Strategies account credential to log in before serving.
Maha MCP Bridge is a security focused tool that lets an AI agent, like Claude Code, work with your local databases and documents without ever sending your raw data to an outside service. It connects using the Model Context Protocol, a standard way for AI tools to plug into external systems, and its whole design is built around the idea that your source data stays on your own machine while only cleaned up, sanitized information reaches the company's API. After installing it globally with npm, you log in with a credential provided by Maha Strategies, which gets checked against their servers once and then stored safely on your computer. From there you run the bridge as a background server that AI clients can connect to, and it refuses to start at all unless you are properly logged in. The bridge exposes a small set of tools an AI agent can call: looking up the structure of a configured local database, running a single read-only SQL query against it, submitting a piece of text for a compliance style audit after redacting sensitive details, and pulling summary counts from a local audit log. Anything read from a database or document passes through two layers of personal information redaction before an agent ever sees it. The first layer uses pattern matching to catch things like emails, phone numbers, and payment card numbers. The second layer runs a small AI model entirely on your own machine to catch names of people, places, and organizations that plain patterns would miss. If that local model cannot load for any reason, the bridge simply falls back to the pattern based redaction instead of failing. Every database query is restricted to read only statements, and the software double checks this at the database engine level too, so nothing can slip through and modify data. All tool activity gets logged locally in a tamper evident record that stores only hashes and counts, never the actual data or queries. The project also includes a separate feature for mounting a purchased book so an AI agent can search and read it section by section instead of scrolling through raw text. The bridge itself is written in TypeScript, MIT licensed, and open source, while the company's broader auditing and billing service sits behind it as a paid product. It currently has 1 star on GitHub.
A local security bridge that lets AI agents query your databases and documents while automatically redacting personal information before anything reaches an external API.
Mainly TypeScript. The stack also includes TypeScript, MCP, ONNX.
The bridge itself is MIT licensed, so it can be used freely, including commercially, as long as the copyright notice is kept.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.