How plmbr/notebook-intelligence works
Notebook Intelligence is a JupyterLab AI assistant and extensible framework. A user opens the chat panel, configures a provider or agent mode, and sends notebook-related requests; the product also offers inline completion, code generation, notebook execution, and management/diagnostic features. The frontend communicates through its API layer with Python services, which coordinate model providers and agent/tool integrations. The README establishes the product capabilities; source-index references support several internal dependencies, but unsampled implementations and route wiring limit which connections can be shown confidently.
Main entry points: ai_service_manager.py AI service manager · api.ts Frontend API · api.py Server API · index.ts Lab extension
Components 22
JupyterLab experience
- Lab extension
src/index.ts - Chat interface
src/chat-sidebar.tsx - Frontend API
src/api.ts - Settings panels
src/components/settings-panel.tsx - First-run tour
src/tour/tour-steps.ts - Chatbook tools
src/chatbook.ts
Assistant services
- Server API
notebook_intelligence/api.py - AI service manager
notebook_intelligence/ai_service_manager.py - LLM providers
notebook_intelligence/llm_providers - Inline completion
notebook_intelligence/inline_completion.py - Chat participants
notebook_intelligence/base_chat_participant.py - Chatbook kernel
notebook_intelligence/chatbook_kernel/kernel.py
Agents and tools
- Claude agent
notebook_intelligence/claude.py - ACP agent
notebook_intelligence/acp_agent.py - MCP tools
notebook_intelligence/mcp_manager.py - Rulesets
notebook_intelligence/rule_manager.py - Claude skills
notebook_intelligence/skill_manager.py - Claude plugins
notebook_intelligence/plugin_manager.py
Configuration and operations
- Configuration
notebook_intelligence/config.py - Performance diagnostics
notebook_intelligence/perf.py - Translations
notebook_intelligence/i18n.py
Other components
- Notebook user
Connections 24
- Notebook user → Chat interface: sends requests
README.md - Lab extension → Chat interface: registers panel
src/index.ts - Lab extension → Frontend API: initializes
src/index.ts - Lab extension → First-run tour: dispatches tour
src/index.ts - Chat interface → Frontend API: uses API
- Settings panels → Frontend API: configures features
src/components/settings-panel.tsx - Frontend API → Server API: sends requests
- Server API → AI service manager: uses services
notebook_intelligence/api.py - AI service manager → LLM providers: selects models
notebook_intelligence/ai_service_manager.py - AI service manager → Claude agent: selects agent mode
notebook_intelligence/ai_service_manager.py - AI service manager → ACP agent: selects agent mode
notebook_intelligence/ai_service_manager.py - Chat participants → Server API: uses chat contracts
notebook_intelligence/base_chat_participant.py - Chat participants → Rulesets: injects rules
notebook_intelligence/base_chat_participant.py - Chat participants → MCP tools: uses toolsets
notebook_intelligence/mcp_manager.py - Claude agent → MCP tools: uses servers
- Claude agent → Claude skills: invokes skills
- Claude agent → Claude plugins: supports plugins
- Frontend API → Chatbook kernel: requests execution
- Lab extension → Chatbook tools: summarizes cells
src/index.ts - AI service manager → Inline completion: selects completion model
- Server API → Configuration: reads settings
notebook_intelligence/api.py - AI service manager → Performance diagnostics: uses diagnostics
notebook_intelligence/ai_service_manager.py - Server API → Translations: uses translations
notebook_intelligence/api.py - Frontend API → Translations: formats interface
src/api.ts