How tanushvasupalli/va works
Wednesday is a voice and remote-control AI assistant, with both a local microphone loop and a Telegram control center. Users speak or send text/voice; the assistant transcribes and interprets input, responds with text or speech, and can invoke memory, automation, device, and PC capabilities. A web dashboard provides another interaction and management surface. The graph preserves documented capabilities, but several tool implementations and runtime routes were not excerpted, so uncertain links are omitted or marked dashed.
Main entry points: brain.py Reasoning Brain · main.py Voice Loop · telegram_bot.py Telegram Bot · mcp_client.py MCP Client
Components 27
User Interfaces
- Telegram Bot
core/telegram_bot.py - Web Dashboard
dashboard/app.py
Voice Processing
- Voice Loop
main.py - Microphone Audio
core/audio_io.py - Speech Segmentation
core/vad.py - Wake Word
core/wake_word.py - Transcription
core/transcriber.py - Speaker Recognition
core/speaker_recognition.py - Speech Output
core/speaker.py
Assistant Intelligence
- Reasoning Brain
core/brain.py - Security Controls
core/security.py
Tools and Integrations
- n8n Workflows
tools/n8n_tools.py - MCP Client
core/mcp_client.py - PC Access
tools/pc_bridge_tools.py - PC Companion
pc_companion.py - Network Tools
tools/network_tools.py - System Tools
tools/system_tools.py - Web, Files, Vision
tools
State and Configuration
- Memory Tools
tools/memory_tools.py - Conversation Storage
core/storage.py - Runtime Configuration
config.py - Config Manager
core/config_manager.py
Other components
- User
- Supabase
- Gemini / Groq
- MCP Servers
- n8n Service
Connections 30
- User → Telegram Bot: sends messages
README.md - User → Voice Loop: speaks to
README.md - User → Web Dashboard: uses
README.md - Voice Loop → Microphone Audio: starts capture
main.py - Microphone Audio → Speech Segmentation: provides chunks
core/vad.py - Voice Loop → Speech Segmentation: records speech
main.py - Voice Loop → Wake Word: checks wake phrase
main.py - Voice Loop → Transcription: transcribes audio
main.py - Voice Loop → Speaker Recognition: identifies speaker
main.py - Voice Loop → Reasoning Brain: submits query
main.py - Reasoning Brain → Speech Output: returns spoken reply
main.py - Telegram Bot → Reasoning Brain: submits requests
README.md - Telegram Bot → Speech Output: sends audio replies
README.md - Web Dashboard → Reasoning Brain: submits text
dashboard/app.py - Reasoning Brain → Gemini / Groq: queries models
core/brain.py - Reasoning Brain → Conversation Storage: reads and writes
core/brain.py - Memory Tools → Conversation Storage: accesses memories
tools/memory_tools.py - Conversation Storage → Supabase: persists data
core/storage.py - Reasoning Brain → n8n Workflows: dispatches workflows
core/brain.py - n8n Workflows → n8n Service: calls workflows
README.md - Reasoning Brain → MCP Client: dispatches MCP tools
core/brain.py - MCP Client → MCP Servers: discovers and calls
core/mcp_client.py - Reasoning Brain → PC Access: dispatches PC actions
core/brain.py - PC Access → PC Companion: requests PC access
- Reasoning Brain → Network Tools: dispatches network actions
core/brain.py - Reasoning Brain → System Tools: dispatches system actions
core/brain.py - Reasoning Brain → Web, Files, Vision: uses assistant tools
core/brain.py - Reasoning Brain → Security Controls: verifies protected actions
core/brain.py - Config Manager → MCP Client: refreshes tool discovery
core/config_manager.py - Config Manager → Runtime Configuration: updates settings
core/config_manager.py