How volcengine/minecontext works
MineContext is a local-first desktop context assistant. Users enable screen recording; captured and other collected context is processed, stored, and later used for proactive summaries, activities, tips, and todos, or resurfaced through AI chat. The graph combines documented product stages with connections visible in the provided source index and excerpts. Several backend routes and frontend service links were not sampled, so those relationships are omitted rather than inferred.
Main entry points: App.tsx Desktop UI · global_storage.py Storage facade · global_vlm_client.py VLM client · smart_todo_manager.py Proactive generation
Components 26
Desktop experience
- Desktop UI
frontend/src/renderer/src/App.tsx - Settings
frontend/src/renderer/src/pages/settings/settings.tsx - Screen monitor
frontend/src/renderer/src/pages/screen-monitor/screen-monitor.tsx - Home dashboard
frontend/src/renderer/src/pages/home/home-page.tsx - AI chat
frontend/src/renderer/src/components/ai-assistant/index.tsx - Vault browser
frontend/src/renderer/src/pages/vault/Vault.tsx
Context collection
- Screen capture
frontend/src/main/background/task/screen-monitor-task.ts - Capture sources
frontend/src/main/utils/get-capture-sources.ts
Context lifecycle
- Capture manager
opencontext/managers/capture_manager.py - Processor manager
opencontext/managers/processor_manager.py - Context processing
opencontext/context_processing/merger/context_merger.py - Storage facade
opencontext/storage/global_storage.py - Unified storage
opencontext/storage/unified_storage.py - SQLite backend
opencontext/storage/backends/sqlite_backend.py
AI and generation
- Consumption scheduler
opencontext/managers/consumption_manager.py - Proactive generation
opencontext/context_consumption/generation/smart_todo_manager.py - Context agent
opencontext/context_consumption/context_agent/agent.py - Tool executor
opencontext/tools/tools_executor.py - Completion service
opencontext/context_consumption/completion/completion_service.py
Shared platform
- Context models
opencontext/models/context.py - VLM client
opencontext/llm/global_vlm_client.py - Embedding client
opencontext/llm/global_embedding_client.py - LLM client
opencontext/llm/llm_client.py - Configuration
opencontext/config/global_config.py - Prompt manager
opencontext/config/prompt_manager.py
Other components
- User
Connections 30
- User → Desktop UI: uses
README.md - Desktop UI → Settings: routes to
frontend/src/renderer/src/App.tsx - Desktop UI → Screen monitor: routes to
- Desktop UI → Home dashboard: routes to
- Desktop UI → AI chat: opens
- Desktop UI → Vault browser: routes to
- User → Screen monitor: starts recording
README.md - Screen monitor → Screen capture: controls
- Screen capture → Capture sources: gets sources
- Capture manager → Processor manager: dispatches context
opencontext/managers/capture_manager.py - Processor manager → Context processing: processes context
- Context processing → Context models: creates context
opencontext/context_processing/merger/context_merger.py - Context processing → Storage facade: stores context
opencontext/context_processing/merger/context_merger.py - Storage facade → Unified storage: delegates storage
opencontext/storage/global_storage.py - Unified storage → SQLite backend: dispatches storage
opencontext/storage/unified_storage.py - Consumption scheduler → Proactive generation: schedules generation
opencontext/managers/consumption_manager.py - Proactive generation → Storage facade: retrieves context
opencontext/context_consumption/generation/smart_todo_manager.py - Proactive generation → VLM client: generates insights
opencontext/context_consumption/generation/smart_todo_manager.py - Proactive generation → Home dashboard: delivers summaries and tasks
- AI chat → Context agent: submits questions
- Context agent → Tool executor: dispatches tools
- Context agent → VLM client: requests responses
- Completion service → VLM client: requests completion
opencontext/context_consumption/completion/completion_service.py - Completion service → Storage facade: reads context
opencontext/context_consumption/completion/completion_service.py - VLM client → LLM client: uses model client
opencontext/llm/global_vlm_client.py - Embedding client → LLM client: uses model client
opencontext/llm/global_embedding_client.py - Storage facade → Embedding client: vectorizes context
- Configuration → Prompt manager: configures prompts
opencontext/config/global_config.py - VLM client → Configuration: loads model config
opencontext/llm/global_vlm_client.py - Embedding client → Configuration: loads model config
opencontext/llm/global_embedding_client.py