RepoMind Bridge
Enables coding agents to receive and act on voice-captured tasks delivered as GitHub issues, including retrieving full context, adding notes, asking the developer questions, and closing tasks when complete.
README
RepoMind Bridge
An MCP server that hands your coding agent the tasks you dictated — in the words you actually said, not a tidied-up title.
You find a bug on your phone, away from the keyboard. You say it out loud. Later you open the repository and Claude Code, Cursor or Codex already knows what you asked for, picks it up, and closes it when the work is done.
phone (dictate) ──► RepoMind ──► GitHub issue ──► this bridge ──► your agent
Tasks travel as GitHub issues labelled repomind-task. There is no server to deploy and
nothing to pay for: GitHub is the postbox.
Install
git clone https://github.com/idanidev/repomind-bridge.git
cd repomind-bridge
npm install
npm run build
Authentication
No token needs to be written into a config file. If you are already signed in with the GitHub CLI, the bridge uses those credentials:
gh auth login
Otherwise set GITHUB_TOKEN to a token with repo scope:
export GITHUB_TOKEN=...
GITHUB_TOKEN wins when both are present. If GitHub rejects it, the bridge resolves it again and
retries once — an MCP server stays up for days, and gh rotates its token underneath it.
Point your agent at it
Claude Code — add to ~/.claude.json:
{
"mcpServers": {
"repomind": {
"command": "node",
"args": ["/absolute/path/to/repomind-bridge/dist/index.js"]
}
}
}
Cursor — the same block in .cursor/mcp.json. Codex — in .codex/config.toml.
Tools
| Tool | What it does |
|---|---|
list_tasks |
Everything captured for this repository. Call it before asking what to work on. |
get_task |
One task in full, including rawInput and the comment thread. |
start_task |
Marks it in progress — visible on the phone — and returns the full context. |
add_note |
Posts a note to the thread, which reaches the developer's phone. |
ask_user |
Marks it blocked and asks one specific question. Do not wait for the answer. |
complete_task |
Closes the issue. Attach a PR url if there is one. |
get_task_image |
The screenshot attached when the task was captured. |
rawInput is the specification
title is a cleaned-up label. rawInput is what the developer dictated. When the two seem to
disagree, rawInput wins — it is the only record of what they meant.
Screenshots never leave the machine
get_task_image reads the image from the RepoMind app's local database, not from GitHub.
Committing screenshots into the repository would grow it permanently, and in a private repo the image would not even render: GitHub does not authenticate image URLs on the reader's behalf. Making the repo public to fix that would publish bug screenshots, which routinely contain personal data.
So nothing is uploaded anywhere. The image is already on the machine, synced by iCloud; the bridge reads it and hands it over. It follows that this tool only works where the Mac app is installed.
Session briefing
src/brief.ts can run as a SessionStart hook so the agent opens already knowing what is waiting.
It stays silent when there is no token, no network, or no tasks — a hook that fails loudly on a
plane is worse than no hook.
Requires
The RepoMind iOS/macOS app to capture the tasks, and a GitHub account. Node 18+.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.