Sandbox
@lincotalk/linco-bridge

Bridge for local agent CLIs and remote clients

Linco Bridge connects agent CLIs running on a personal computer to web, H5, mini program, app, and IM clients. The local connector handles sessions, permissions, attachments, and generated files, while the platform and docs provide a reference backend, web UI, and protocol guidance.

90 stars7 forksJavaScriptUpdated 8d ago
Who it's for

Builders who want to control local agent sessions from web, mobile, or chat clients.

What it delivers

You can keep agent execution local while still viewing, messaging, and managing sessions from remote clients.

What it does

Local connector for agent CLIs

Runs `linco-connect` on the user's computer and connects it to an agent CLI session.

WebSocket bridge

Uses an authenticated WebSocket connection between the local connector and the remote channel.

Reference platform

Provides a NestJS backend and UniApp web client for the open `linco-demo` channel.

Custom channel adapters

Lets you build new H5, mini program, app, web, or IM clients through the connector and protocol helpers.

Session and file access

Exposes sessions, permissions, attachments, and generated files to remote clients without moving execution off the machine.

How to get it

  1. 1Linco Bridge connects local Agent CLIs to remote clients without moving Agent execution…
    Local Agent CLI
        ↕ local process, gateway, or session files
    linco-connect on the user's computer
        ↕ authenticated WebSocket bridge
    Official Linco channel, open reference platform, or custom backend
        ↕
    Linco App, Reference Web, or custom client
  2. 2Open Bridge, choose an Agent import path, and run the generated commands on the computer…
    npm install -g linco-connect
    linco-connect init --token "<app-id>:<app-secret>" --agent codex
    linco-connect start --daemon
  3. 3Start the backend
    cd linco-bridge-platform/server
    npm install
    npm run start:dev
  4. 4Start the H5 frontend in another terminal
    cd linco-bridge-platform/web
    npm install
    npm run dev:h5
  5. 5The generated commands are the source of truth. A local command usually has this shape
    linco-connect init \
      --token "<app-id>:<app-secret>" \
      --agent codex \
      --channel linco-demo \
      --account "<account-id>" \
      --allow-insecure-ws
    
    linco-connect start --daemon

README

Linco Bridge logo

Linco Bridge

An open bridge layer that connects AI Agent tools running on a personal computer to web, H5, mini program, app, IM, and custom clients.

status: open source alpha supported agents channel types license: MIT

Status: Open Source Alpha. Interfaces and compatibility may change with documented migration notes.

iOS App · Android App · Hosted Demo · Watch Demo · Quick Start · 简体中文

Linco Bridge hero preview

✨ Highlights

Local-firstAgent-oriented interactionOpen extension surface
Keep Agent CLIs running on the user's computer while accessing sessions, attachments, and generated files from remote clients.Present streaming output, permissions, tools, files, and long-running sessions in a dedicated client instead of a generic message relay.Reuse the connector, protocol helpers, SDKs, and channel adapters to build a custom H5, mini program, app, web, or IM experience.

🧭 Overview

Linco Bridge connects local Agent CLIs to remote clients without moving Agent execution away from the user's computer.

Local Agent CLI
    ↕ local process, gateway, or session files
linco-connect on the user's computer
    ↕ authenticated WebSocket bridge
Official Linco channel, open reference platform, or custom backend
    ↕
Linco App, Reference Web, or custom client
Repository areaPurpose
linco-bridge-connectLocal connector for Agent CLIs, sessions, permissions, attachments, and generated files.
linco-bridge-platformNestJS + UniApp reference implementation for the open linco-demo channel.
docsSetup, architecture, protocol, security, compatibility, and extension guides.

The repository does not include the complete Linco App or the official hosted cloud-service implementation.

🚀 Choose a Path

PathBest forRequirement
Linco AppUsers who want the official product experienceInstall the app and run the page-generated connector commands.
Open reference platformLocal validation, self-hosting evaluation, and secondary developmentRun linco-bridge-platform/server and linco-bridge-platform/web.
Hosted demoQuick public evaluation without deploying the platformOpen the hosted H5 or mini program and run the generated connector commands locally.

1. Linco App

Download Linco App:

Open Bridge, choose an Agent import path, and run the generated commands on the computer where the Agent CLI is installed. A typical command shape is:

npm install -g linco-connect
linco-connect init --token "<app-id>:<app-secret>" --agent codex
linco-connect start --daemon

Return to the app, confirm the connector is online, and send a test message.

2. Open Reference Platform

Start the backend:

cd linco-bridge-platform/server
npm install
npm run start:dev

Start the H5 frontend in another terminal:

cd linco-bridge-platform/web
npm install
npm run dev:h5

Open the printed H5 URL, then:

  1. Open Bridge and select Import from Codex.
  2. Copy the generated setupCommands and run them locally.
  3. Return to the page and refresh the connection status.
  4. Enter Codex, select a project or session from the folder menu, and send a test message.

The generated commands are the source of truth. A local command usually has this shape:

linco-connect init \
  --token "<app-id>:<app-secret>" \
  --agent codex \
  --channel linco-demo \
  --account "<account-id>" \
  --allow-insecure-ws

linco-connect start --daemon

See Quick Start and the Platform README for the complete local flow.

3. Hosted Demo

Linco Bridge WeChat Mini Program QR code

Open Bridge, import an Agent, run the generated setupCommands on your computer, confirm the connector is online, and enter the chat page.

The hosted demo uses signed anonymous visitor sessions. It does not provide a formal account, durable cross-device recovery, or production-grade multi-tenant guarantees. Do not enter sensitive or production data. Custom frontends must self-host the platform instead of calling the official demo API.

🤖 Compatibility

AgentStatusVerified version
Codex CLISupportedcodex-cli 0.142.5
Claude CodeSupported2.1.198 (Claude Code)
HermesSupportedHermes Agent v0.13.0 (2026.5.7)
OpenClawSupportedOpenClaw 2026.5.18 (50a2481)

Use Node.js 20 or 22–26; Node.js 22 LTS is recommended. See Supported Platforms for the current compatibility boundary.

🧩 Extension Points

AreaLocation
Connector SDKlinco-bridge-connect/src/package/connector
Protocol helperslinco-bridge-connect/src/package/protocol
Reference Web SDKslinco-bridge-platform/web/src/bridge/sdk
Channel adapterslinco-bridge-connect/src/channel

Third-party integrations should add a custom channel adapter rather than changing the official linco channel. See Secondary Development Rules.

🔐 Security

  • Never commit or log tokens, App Secrets, private keys, user data, or authenticated URLs.
  • Use ws:// only for trusted local development. Public deployments should use HTTPS/WSS and define their own authentication, storage, audit, and retention policies.
  • TLS/WSS does not by itself provide end-to-end encryption.
  • File delivery must keep hidden and sensitive paths blocked.
  • Read Security and Privacy before connecting real data. Report vulnerabilities through SECURITY.md.

📚 Documentation

GoalDocumentation
Get startedQuick Start, Troubleshooting
Understand the systemHow It Works, Protocol, CLI Reference
Run or extend the platformHosted Deployment, Reference Web, Secondary Development
Review scope and policySupported Platforms, Security and Privacy, Support
Work on a subprojectConnector README, Platform README

💬 Community

Join the WeChat technical group for integration questions, product feedback, and Agent workflow discussions:

Linco technical WeChat group

Follow Linco Lab for project updates and implementation notes:

  • Xiaohongshu
  • WeChat Official Account: search Linco Lab
Linco Lab WeChat Official Account QR code

WeChat QR codes may expire. See Community for the latest entries and support guidance.

🤝 Contributing

Issues and pull requests are welcome. Read CONTRIBUTING.md before proposing protocol, cross-module, or breaking changes.

⚖️ License

Linco Bridge is licensed under the MIT License. The license applies to repository code and documentation; it does not grant separate rights to use Linco names, logos, or other brand assets beyond normal nominative reference.

Files in the repo

Repository payload14 top-level entries
  • .github
  • docs
  • linco-bridge-connect
  • linco-bridge-platform
  • CHANGELOG.md
  • CODE_OF_CONDUCT.md
  • COMMUNITY.md
  • COMMUNITY.zh-CN.md
  • CONTRIBUTING.md
  • LICENSE
  • README.md
  • README.zh-CN.md
  • SECURITY.md
  • SUPPORT.md

Discussion (0)

Ask about usage, or say what you built with it

Sign in to join the discussion.

No comments yet. Be the first to say what this is good for.

More connectors

Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface

86k

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

43k

Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code

14k
okf-memory/
okf-agent-memory

Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.

547
tirth8205/
code-review-graph

Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.

31k
2akouwu/
reverify

Stop your AI from making things up — it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is the proving ground. MCP server + CLI.

1.1k