In the modern digital workplace, teams often find themselves navigating a labyrinth of disconnected tools—calendars that don't talk to documents, emails that ignore task lists, and meeting minutes that vanish into the void. Lark Workplace Automation Bridge (formerly lark-cowork-plugin) transforms this chaos into a symphony of seamless productivity. This open-source repository provides a robust API integration layer that connects Claude AI directly with Lark/Feishu's ecosystem, enabling intelligent automation across calendar management, messaging, document creation, spreadsheets, Bitable databases, task tracking, email handling, and meeting minutes.
Think of it as a digital conductor for your enterprise orchestra—Claude AI becomes the maestro that coordinates every instrument in Lark's suite, ensuring no note is missed and every movement flows harmoniously.
- Why This Exists
- Architecture Overview
- Key Features
- Compatibility Matrix
- Installation
- Configuration
- Usage Examples
- API Integration
- Mermaid Diagram
- Example Profile Configuration
- Example Console Invocation
- Troubleshooting
- Contributing
- License
- Disclaimer
The enterprise collaboration landscape is fragmented. Lark/Feishu offers a powerful suite of tools, but without intelligent orchestration, teams spend 30% of their time context-switching between modules. Lark Workplace Automation Bridge emerged from a simple observation: Claude AI can be the bridge that connects these islands of functionality. Instead of manually copying data from a Bitable to a document, or hunting through email threads for meeting action items, this plugin enables Claude to understand your workflow patterns and execute cross-module operations with a single natural language command.
This isn't just another plugin—it's a cognitive extension for your team's collaboration infrastructure.
The plugin operates as a middleware layer between Claude AI's API and Lark's Open Platform. It consists of:
- Core Bridge Engine: Translates natural language commands into Lark API calls
- Module Adapters: Specialized connectors for each Lark module (Calendar, Messages, Docs, Sheets, Bitable, Tasks, Email, Minutes)
- Context Memory: Maintains session state across interactions
- Security Layer: OAuth 2.0 token management and permission scoping
- Natural language scheduling ("Schedule a meeting with the marketing team next Tuesday at 3 PM")
- Automatic conflict resolution across team calendars
- Smart reminder creation based on email content
- Cross-timezone meeting optimization for global teams
- Automated response drafting using Claude's contextual understanding
- Thread summarization for long conversations
- Priority detection and smart notification routing
- Multi-language message composition and translation
- Template-based document generation from meeting notes
- Real-time collaborative editing suggestions
- Automatic formatting and style consistency checks
- Version control with semantic diff analysis
- AI-generated formulas and data transformations
- Natural language query execution on Bitable databases
- Automated data enrichment from external sources
- Visualization recommendations based on data patterns
- Convert email threads into structured task lists
- Automatic task assignment based on email content analysis
- Deadline extraction and calendar integration
- Progress tracking with automated status updates
- Real-time transcription summarization
- Action item detection and assignment
- Decision logging with context preservation
- Minutes distribution with stakeholder tagging
| Operating System | Lark SDK Version | Claude API Version | Status |
|---|---|---|---|
| macOS 14+ | v2.1.0+ | Claude 3.5+ | ✅ Certified |
| Windows 11 | v2.0.0+ | Claude 3.0+ | ✅ Certified |
| Ubuntu 22.04+ | v2.0.0+ | Claude 3.0+ | ✅ Certified |
| CentOS 8+ | v1.9.0+ | Claude 2.1+ | |
| iOS 17+ | v2.1.0+ | Claude 3.5+ | ✅ Mobile |
| Android 14+ | v2.0.0+ | Claude 3.0+ | ✅ Mobile |
- Python 3.10+ or Node.js 18+
- Lark/Feishu developer account with API access
- Active Claude API key (Anthropic)
- Enterprise Lark tenant with admin privileges
# Python installation
pip install lark-cowork-bridge
# Node.js installation
npm install @lark-automation/bridge- Clone the repository:
git clone https://github.com/your-org/lark-cowork-plugin.git
cd lark-cowork-plugin- Install dependencies:
pip install -r requirements.txt
# or
npm install-
Configure environment variables (see Configuration section)
-
Run the initialization script:
python bridge-init.py
# or
node init-bridge.js# Required
LARK_APP_ID=cli_xxxxxxxxxxxxx
LARK_APP_SECRET=xxxxxxxxxxxxxxxxxxxx
CLAUDE_API_KEY=sk-ant-xxxxxxxxxxxxxxxxxxxx
# Optional
LARK_TENANT_ID=your_tenant_id
BRIDGE_LOG_LEVEL=INFO
BRIDGE_CACHE_SIZE=1000The bridge supports dual AI provider architecture for enhanced reliability:
{
"ai_providers": {
"primary": {
"provider": "claude",
"api_key": "${CLAUDE_API_KEY}",
"model": "claude-3-opus-20240229",
"max_tokens": 4096
},
"fallback": {
"provider": "openai",
"api_key": "${OPENAI_API_KEY}",
"model": "gpt-4-turbo",
"max_tokens": 4096
}
},
"routing": "intelligent",
"context_window": 8192
}profiles:
executive_assistant:
name: "Executive Bridge Assistant"
modules:
- calendar
- email
- tasks
permissions:
- calendar:read
- calendar:write
- email:read
- email:send
- tasks:create
ai_config:
system_prompt: "You are an executive assistant managing calendar, email, and tasks."
temperature: 0.3
auto_approve: false
team_automator:
name: "Team Workflow Automator"
modules:
- docs
- sheets
- bitable
- minutes
permissions:
- docs:create
- docs:edit
- sheets:write
- bitable:read
- minutes:generate
ai_config:
system_prompt: "You automate team documentation and data workflows."
temperature: 0.5
auto_approve: true# Schedule a meeting with comprehensive context
python bridge-cli.py --action schedule_meeting \
--title "Q4 Strategy Review" \
--attendees "alice@company.com,bob@company.com" \
--duration 60 \
--preferred-day Tuesday \
--context "Review quarterly metrics and discuss 2026 roadmap"
# Generate meeting minutes from recorded transcript
python bridge-cli.py --action generate_minutes \
--source recording.mp4 \
--template "executive_summary" \
--distribute true \
--stakeholders "leadership@company.com"
# Create Bitable database from spreadsheet
python bridge-cli.py --action spreadsheet_to_bitable \
--source sales_data.xlsx \
--target-db "Sales Pipeline 2026" \
--schema-inference true \
--link-related truegraph TB
User[User/Administrator] --> CLI[Bridge CLI Interface]
User --> API[REST API Endpoint]
User --> Web[Web Dashboard]
CLI --> Bridge[Core Bridge Engine]
API --> Bridge
Web --> Bridge
Bridge --> AI[AI Orchestrator]
Bridge --> Auth[Authentication Layer]
Bridge --> Cache[Context Cache]
AI --> Claude[Claude API]
AI --> OpenAI[OpenAI API]
Auth --> LarkAuth[Lark OAuth 2.0]
subgraph Lark[Lark/Feishu Ecosystem]
Calendar[Calendar API]
Messages[Messages API]
Docs[Docs API]
Sheets[Sheets API]
Bitable[Bitable API]
Tasks[Tasks API]
Email[Email API]
Minutes[Minutes API]
end
Claude --> Lark
OpenAI --> Lark
context((Context Memory))
Cache --> context
event((Event System))
Bridge --> event
event --> Webhooks[Webhook Manager]
Webhooks --> External[External Services]
| Symptom | Cause | Solution |
|---|---|---|
| API timeout errors | Large payload size | Enable streaming mode in configuration |
| Permission denied | Incorrect scope configuration | Review Lark app permissions in developer console |
| Context loss between calls | Cache expiration | Increase BRIDGE_CACHE_SIZE or reduce TTL |
| Rate limiting | Excessive API calls | Implement request queuing with retry logic |
We welcome contributions from the community! Please see our Contributing Guidelines for detailed information on:
- Code style and standards
- Pull request process
- Testing requirements
- Documentation expectations
This project is licensed under the MIT License - see the LICENSE file for details.
Important Notice: Lark Workplace Automation Bridge is an independent open-source project and is not officially affiliated with Lark Technologies Pte. Ltd., Feishu, or Anthropic. While this plugin interfaces with both Lark and Claude APIs, users must comply with the respective terms of service, privacy policies, and data handling requirements of each platform.
Data Privacy: This bridge processes potentially sensitive business communications and calendar data. Deployers are responsible for ensuring compliance with their organization's data governance policies, including:
- GDPR compliance for European operations
- SOC 2 requirements for data handling
- Internal IT security protocols
- Data residency and sovereignty requirements
Limitation of Liability: The maintainers of this project assume no liability for data loss, security breaches, or operational disruptions arising from the use of this bridge. Users should implement proper testing, monitoring, and rollback procedures before production deployment.
AI Output Accuracy: While Claude and OpenAI models strive for accuracy, AI-generated content (meeting minutes, email drafts, task assignments) should be reviewed by humans before acting upon them. The bridge includes confidence scoring for generated content but cannot guarantee 100% accuracy.
Version Compatibility: As Lark/Feishu and Claude APIs evolve, certain features may change or break. We recommend pinning dependency versions and testing before major updates. The bridge includes a compatibility checker that runs at startup.
Built for teams who believe their tools should work as hard as they do. Version 2.0.0 - 2026 Edition