OpenCode Chat Bot

fork from https://github.com/grinev/opencode-telegram-bot

This project differs from the original project in the following ways:

1. OpenClaw plugin — can run as an OpenClaw plugin, leveraging OpenClaw message Channels to support almost all major IM platforms.
2. DingTalk & Feishu standalone — supports both Feishu and DingTalk as independent bot platforms.
3. Telegram removed — Telegram support has been removed. If you need Telegram, use the original project directly.

OpenCode Chat Bot is a secure multi-platform client for OpenCode CLI that runs on your local machine. It supports DingTalk and Feishu as standalone chat bots, and OpenClaw through a plugin entrypoint.

Run AI coding tasks, monitor progress, switch models, and manage sessions from your phone.

For DingTalk and Feishu, no public inbound port is required. The bot communicates with your local OpenCode server and the platform APIs. OpenClaw support is loaded by the OpenClaw plugin runtime.

Scheduled tasks are supported across the shared runtime. OpenClaw integration is intentionally implemented as a thin adapter that reuses the same command, prompt, session, permission, and scheduled-task modules as the standalone bots.

Platforms: macOS, Windows, Linux

Languages: English (en), Deutsch (de), Español (es), Русский (ru), 简体中文 (zh), 繁體中文 (zh-TW)

Demo gif

Features

• Multi-platform support — DingTalk and Feishu standalone bots, plus OpenClaw plugin runtime support
• Shared OpenClaw adapter — OpenClawCode functionality is rebuilt on top of the existing shared core instead of duplicating command or session logic
• Remote coding — send text prompts to OpenCode from chat and receive aggregated results
• Session management — create new sessions or continue existing ones, just like in the TUI
• Live status — pinned message with current project, model, context usage, and changed files list, updated in real time
• Model switching — pick models from OpenCode favorites and recent history directly in the chat (favorites are shown first)
• Agent modes — switch between Plan and Build modes on the fly
• Subagent activity — watch live subagent progress in chat, including the current task, agent, model, and active tool step
• Custom Commands — run OpenCode custom commands (and built-ins like init/review) from an inline menu with confirmation
• Interactive Q&A — answer agent questions and approve permissions through chat interactions
• Permission controls — allow once, always allow, reject, or enable auto-confirmation for a session
• Scheduled tasks — schedule prompts to run later or on a recurring interval; see Scheduled Tasks
• Context control — compact context when it gets too large, right from the chat
• Input flow control — when an interactive flow is active, the bot accepts only relevant input to keep context consistent and avoid accidental actions
• Security — optional DingTalk/Feishu user allowlists restrict access
• Localization — UI localization is supported for multiple languages (BOT_LOCALE)
• File logging — logs are written to files with automatic rotation and retention

Current message input support is text-first. DingTalk and Feishu non-text messages such as images, files, voice, audio, and media are currently reported as unsupported by the platform handlers.

Planned features currently in development are listed in Current Task List.

Prerequisites

• Node.js 20+ — download
• OpenCode — install from opencode.ai or GitHub
• Bot Platform — at least one of:
  • DingTalk Robot (see DingTalk Bot)
  • Feishu Bot (see Feishu Bot)
  • OpenClaw plugin runtime (see OpenClaw Plugin Commands)

Quick Start

1. Choose an Integration

Use one or more of these integration modes:

• DingTalk standalone bot — configure DINGTALK_APP_KEY and DINGTALK_APP_SECRET
• Feishu standalone bot — configure FEISHU_APP_ID and FEISHU_APP_SECRET
• OpenClaw plugin — build this package and let OpenClaw load dist/openclaw-plugin.js

For standalone mode, at least one of DingTalk or Feishu must be configured.

2. Start OpenCode Server

Start the OpenCode server:

opencode serve

The bot connects to the OpenCode API at http://localhost:4096 by default. Override it with OPENCODE_API_URL.

3. Install from npm

The package is published on npm as @luoyingwen/opencode-chat-bot. Choose one of the two modes below — they cannot be used together.

Option A: OpenClaw Plugin (recommended for OpenClaw users)

Advantages over the standalone mode:

• Supports all major IM platforms supported by OpenClaw Channels (Feishu, DingTalk, WeChat, Slack, Discord, Telegram, etc.)
• After entering OpenCode mode via /opencode, all messages are sent directly to OpenCode, bypassing the OpenClaw Agent — faster response, lower cost, and no routing errors
• Exiting OpenCode mode with /exit restores the original OpenClaw conversation flow

openclaw plugins install @luoyingwen/opencode-chat-bot
openclaw config set plugins.entries.opencode-chat-bot.config.enabled true
openclaw gateway restart

Send /opencode in the conversation to enter OpenCode mode, then send prompts or commands.

Option B: Standalone Bot (recommended for DingTalk/Feishu users)

Advantages over the OpenClaw plugin mode:

• Remote coding from your mobile device — send prompts, monitor progress, and manage sessions without sitting at your computer
• No OpenClaw plugin runtime required — works with just DingTalk or Feishu

npm install -g @luoyingwen/opencode-chat-bot
opencode-chat-bot config
opencode-chat-bot start

The config wizard will guide you through selecting a language and configuring Feishu or DingTalk credentials.

For Linux systemd setup, see docs/LINUX_SYSTEMD_SETUP.md.

4. Run From Source

git clone https://github.com/luoyingwen/opencode-chat-bot.git
cd opencode-chat-bot
npm install
cp .env.example .env
# Edit .env with your OpenCode model and platform credentials
npm run dev

5. OpenClaw Local Plugin Install

For local development and testing, build and install the plugin from source:

npm install
npm run build:openclaw
npm run openclaw:install -- local
openclaw config set plugins.entries.opencode-chat-bot.enabled true

For local iteration, use link mode instead:

npm run openclaw:install -- link

6. Installed CLI Mode

When installed globally via npm, use the CLI commands directly:

opencode-chat-bot config
opencode-chat-bot start

On first launch, the interactive wizard asks for interface language and platform credentials.

Supported Platforms (Node.js)

Bot Commands

Shared Commands (All Platforms)

All platforms support the same core command set:

Permission quick replies (all platforms):

Any regular text message is sent as a prompt to the coding agent only when no blocking interaction is active.

Platform-Specific Commands

Note: DingTalk and Feishu currently support text and markdown messages. Image, voice, and file messages will show a "not supported" notice.

Under normal usage, start opencode serve yourself before launching the standalone bot. OpenClaw users should send /opencode in the conversation before sending OpenCode commands or prompts.

OpenClaw Plugin Setup

OpenClaw loads ./dist/openclaw-plugin.js and reuses the same shared command, prompt, task, permission, and route-scoped state modules as DingTalk and Feishu. Use /opencode in a conversation to enter OpenCode mode. While active, OpenClaw commands and normal text are handled by OpenCode. Use /exit to leave OpenCode mode for that conversation.

Build the OpenClaw plugin entrypoint with either command below. The original OpenClawCode build script was tsc; in this merged repository the normal build compiles the standalone bot and the OpenClaw plugin together.

npm run build
# or, for an explicit plugin-oriented alias:
npm run build:openclaw

For local OpenClaw testing, install the built plugin with the migrated install script:

npm run openclaw:install -- local

# or link this working tree while iterating:
npm run openclaw:install -- link

After building, OpenClaw can discover the plugin from package.json under openclaw.extensions, from the ./openclaw-plugin export, or from openclaw.plugin.json for runtimes that read plugin metadata files.

Example OpenClaw configuration:

openclaw config set plugins.entries.opencode-chat-bot.enabled true

DingTalk Bot Setup

DingTalk uses Stream Mode (no webhook server required) for real-time message reception.

Step 1: Create a DingTalk Robot

1. Log in to DingTalk Developer Platform
2. Create an Enterprise Internal Application (企业内部应用)
3. Go to Robot & Message (机器人与消息推送) section
4. Create a Stream Mode Robot (Stream 模式机器人)

Step 2: Get Credentials

From the application details page:

• AppKey → DINGTALK_APP_KEY
• AppSecret → DINGTALK_APP_SECRET

Step 3: Configure User Access

Bot automatically locks to the first user who sends a message.

• If unset: Auto-lock to first user
• If set: Only matching user can access

Set DINGTALK_ALLOWED_USER_ID to pre-configure a specific DingTalk staff ID.

Step 4: Configure Environment

Add to your .env:

DINGTALK_APP_KEY=your-app-key
DINGTALK_APP_SECRET=your-app-secret
# DINGTALK_ALLOWED_USER_ID= (optional, auto-locks to first user if unset)

Step 5: Test

1. Start the bot: npm run dev
2. Find the robot in DingTalk and send a message
3. Use /status to verify connection

Proactive Messaging

The bot supports proactive messaging via DingTalk's oToMessages/batchSend API. This allows:

• Scheduled task notifications — receive alerts without sending a message first
• Push notifications — bot can message you proactively

Permission Requirements:

In DingTalk Developer Platform, ensure your app has:

• ChatBot.SendMessage — Send bot messages to users

If permissions are missing, the bot will fall back gracefully and log a warning.

Note: DingTalk currently supports text and markdown messages. Image, voice, and file messages will show a "not supported" notice.

Feishu Bot

Feishu uses Stream Mode for real-time message reception.

Step 1: Create a Feishu Bot

1. Log in to Feishu Developer Platform
2. Create a Custom App (自定义应用)
3. Go to Robot (机器人) section and enable robot capability

Step 2: Get Credentials

From the application details page:

• App ID → FEISHU_APP_ID
• App Secret → FEISHU_APP_SECRET

Step 3: Configure User Access

Bot automatically locks to the first user who sends a message.

• If unset: Auto-lock to first user
• If set: Only matching user can access

Set FEISHU_ALLOWED_USER_ID to pre-configure a specific Feishu user ID (open ID).

Step 4: Configure Environment

Add to your .env:

FEISHU_APP_ID=your-app-id
FEISHU_APP_SECRET=your-app-secret
# FEISHU_ALLOWED_USER_ID= (optional, auto-locks to first user if unset)

Step 5: Test

1. Start the bot: npm run dev
2. Find the bot in Feishu and send a message
3. Use /status to verify connection

Note: Feishu currently supports text and markdown-style post content. Image, file, audio, and media messages show a "not supported" notice.

Scheduled Tasks

Scheduled tasks let you prepare prompts in advance and run them automatically later or on a recurring schedule. This is useful for periodic checks, routine code maintenance, or tasks you want OpenCode to execute while you are away from your computer. Use /task to create a scheduled task and /tasks to review or delete existing ones.

• Each task is created from the currently selected OpenCode project and model
• Scheduled executions currently always run with the build agent
• Tasks run outside your active chat session, so they do not interrupt or affect the current session flow
• The minimum recurring interval is 5 minutes
• Up to 10 scheduled tasks can exist at once by default; change this with TASK_LIMIT in your .env

Configuration

Localization

• Supported locales: en, de, es, ru, zh, zh-TW
• The setup wizard asks for language first
• You can change locale later with BOT_LOCALE

Environment Variables

When installed via npm, the configuration wizard handles the initial setup. The .env file is stored in your platform's app data directory:

• macOS: ~/Library/Application Support/opencode-chat-bot/.env
• Windows: %APPDATA%\opencode-chat-bot\.env
• Linux: ~/.config/opencode-chat-bot/.env

*At least one standalone bot platform must be configured when running opencode-chat-bot start: DingTalk (DINGTALK_APP_KEY + DINGTALK_APP_SECRET) or Feishu (FEISHU_APP_ID + FEISHU_APP_SECRET). OpenClaw is loaded separately by the OpenClaw plugin runtime.

Keep your .env file private. It contains your bot tokens. Never commit it to version control.

Logs are written to ./logs when running from sources and to the runtime config directory logs/ folder in installed mode. Log rotation depends on runtime mode: sources creates one file per bot launch, while installed appends to one file per day. Old log files are removed according to LOG_RETENTION.

Model Configuration

The model picker uses OpenCode local model state (favorite + recent):

• Favorites are shown first, then recent
• Models already in favorites are not duplicated in recent
• Current model is marked with ✅
• Default model from OPENCODE_MODEL_PROVIDER + OPENCODE_MODEL_ID is always included in favorites
• Use /models to list available models and /model <n> to select one
• Model selection is route-scoped and affects future prompts; in-flight responses are not interrupted

To add a model to favorites, open OpenCode TUI (opencode), go to model selection, and press Cmd+F/Ctrl+F on the model.

Security

The bot enforces strict user whitelists:

• DingTalk: Auto-locked to first user, or matches DINGTALK_ALLOWED_USER_ID if pre-configured
• Feishu: Auto-locked to first user, or matches FEISHU_ALLOWED_USER_ID if pre-configured

Messages from unauthorized sources are silently ignored and logged.

Since the bot runs locally on your machine and connects to your local OpenCode server, there is no external attack surface beyond the platform APIs.

Available Scripts

Note: No file watcher or auto-restart is used. The bot maintains persistent SSE and long-polling connections — automatic restarts would break them mid-task. After making changes, restart manually with npm run dev.

Troubleshooting

Bot doesn't respond to messages

• Make sure user/channel ID matches your actual ID
• Verify the bot token/credentials are correct

"OpenCode server is not available"

• Ensure opencode serve is running in your project directory
• Check that OPENCODE_API_URL points to the correct address (default: http://localhost:4096)

No models in model picker

• Add models to your OpenCode favorites: open OpenCode TUI, go to model selection, press Ctrl+F on desired models
• Verify OPENCODE_MODEL_PROVIDER and OPENCODE_MODEL_ID point to an available model in your setup

DingTalk not receiving messages

• Verify robot is created with Stream Mode enabled
• Check DINGTALK_APP_KEY and DINGTALK_APP_SECRET are correct
• Ensure the robot is published and available to users

Feishu not receiving messages

• Verify bot is enabled and webhook URL is configured correctly
• Check FEISHU_APP_ID and FEISHU_APP_SECRET are correct
• Check FEISHU_ALLOWED_USER_ID if access is restricted
• Ensure the bot is published and available to users

OpenClaw plugin does not respond

• Build first with npm run build:openclaw
• Install or link locally with npm run openclaw:install -- local or npm run openclaw:install -- link
• Enable the plugin with openclaw config set plugins.entries.opencode-chat-bot.enabled true
• Send /opencode in the conversation before sending OpenCode commands or prompts
• Check scope filters in channels, accountIds, and conversationIds

Linux: permission denied errors

• Make sure the CLI binary has execute permission: chmod +x $(which opencode-chat-bot)
• Check that the config directory is writable: ~/.config/opencode-chat-bot/

Contributing

Please follow commit and release note conventions in CONTRIBUTING.md.

Community

Have questions, want to share your experience using the bot, or have an idea for a feature? Join the conversation in GitHub Discussions.

License

MIT © Tony Lee