T3 Bridge
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@T3 Bridgelist all tools from my T3 instances"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
T3 Bridge
Connect multiple T3 Code MCP servers through one authenticated endpoint.
T3 Bridge is intended for people who run more than one T3 Code environment and want one MCP connection for clients such as ChatGPT. Each instance's tools will have a distinct prefix so clients can tell them apart.
Project status
This repository contains the initial project scaffold. The gateway implementation and installation tooling have not been migrated yet. There is no runnable bridge or installer in this revision.
T3 Bridge is an independent community project and is not affiliated with or endorsed by T3 Code, OpenAI, or ngrok.
Related MCP server: mcp-gway
Intended scope
One authenticated MCP endpoint for multiple T3 Code environments.
Separate OAuth authorization for each upstream instance.
Instance prefixes on tool names to avoid collisions.
One owner account, configured through a file.
Persistent OAuth credentials stored outside the repository.
Deployment behind a stable HTTPS endpoint, including ngrok.
T3 remains responsible for its own permissions. A bridge connection must never turn a read-only T3 grant into write access. Each upstream must explicitly approve the actions its tools perform.
Development
The scaffold uses Node.js 24, pnpm 11.18.0, and TypeScript. The pnpm version is pinned in package.json; see the pnpm installation guide if it is not installed.
pnpm install --frozen-lockfile
pnpm run check
pnpm run buildThese commands check and compile the TypeScript scaffold. Automated tests and an application start command will be added with the implementation.
See CONTRIBUTING.md for contribution guidance and SECURITY.md for reporting vulnerabilities.
Configuration example
examples/config.example.json illustrates the intended configuration using documentation-only hostnames and an example email. It is a design example; the scaffold does not load it yet.
Use local configuration files for real addresses. Keep passwords, ngrok tokens, OAuth credentials, signing keys, databases, and authorization URLs out of commits and issue reports. The repository's ignore rules exclude common local configuration and state files, but they do not replace reviewing a diff before publishing it.
License
MIT.
This server cannot be deployed
Maintenance
Related MCP Connectors
One OAuth connector for your team's remote MCP servers, with Tool Packs, guardrails and logs.
- 0bridgeOAuthdev.0bridge
Every service you connect, one MCP endpoint for all your AI tools. Sign in once.
One MCP endpoint for your company's tools, with per-person sign-in, policy and audit.
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables clients to access multiple backend MCP servers through a single endpoint, with OAuth 2.1 authorization, namespaced tools, and secure credential management.2MIT
- AlicenseNot gradedqualityAmaintenanceEnables aggregating multiple MCP servers behind a single HTTP/SSE endpoint, with Code Mode to reduce token usage and support for OAuth and sandboxed execution.807 npmMIT
- AlicenseNot gradedqualityBmaintenanceAggregates multiple MCP servers behind a single endpoint, letting MCP clients access all their tools through one URL and token while namespacing tools to avoid collisions.MIT
- AlicenseNot gradedqualityBmaintenanceEnables connecting multiple remote HTTP MCP servers through a single local stdio endpoint, handling OAuth authentication, token refresh, and namespace-scoped tools.20 npmMIT