How 17rajsal/boloai works
BoloAI is a voice-first digital-services gateway: callers use an ordinary phone or the web interface, speak a request, and receive a concise spoken result. Its main call flow transcribes audio, reasons over session context, uses domain tools, verifies results, and speaks the response; confirmed actions can dispatch SMS or register grievances. A separate upload handoff analyzes submitted documents. The README documents several integrations and dashboard behaviors beyond the sampled implementation, so those links are shown as documented rather than independently verified in source.
Main entry points: orchestrator.py Agent orchestration · main.py FastAPI backend · registry.py Tool registry · verification.py Source verification
Components 27
Caller access
- Exotel PSTN
- FastAPI backend
app/main.py
Voice conversation
- Speech recognition
app/adapters/stt.py - Agent orchestration
app/agent/orchestrator.py - Session context
app/agent/memory.py - Voice guidance
app/agent/prompt.py - Speech synthesis
app/adapters/tts.py - Audio stream adapter
app/adapters/telephony.py
Safety and verification
- Safety guardrails
app/services/safety.py - Source verification
app/services/verification.py - Action lifecycle
app/services/actions.py
Digital services
- Tool registry
app/tools/registry.py - Weather lookup
app/tools/weather.py - Government schemes
app/tools/schemes.py - Courier tracking
app/tools/courier.py - Web search
app/tools/search.py - SMS dispatch
app/tools/sms.py - Document handoff
app/services/upload.py - Image analysis
app/adapters/multimodal.py
Dashboard and state
- Session and events
app/store.py - Live dashboard
app/static/app.js
Other components
- Phone caller
- Web user
- Open-Meteo
- Sarvam AI
- AI model provider
- Official web sources
Connections 30
- Phone caller → Exotel PSTN: calls
README.md - Exotel PSTN → FastAPI backend: streams audio
README.md - FastAPI backend → Audio stream adapter: handles stream
README.md - FastAPI backend → Speech recognition: transcribes audio
app/main.py - Speech recognition → Sarvam AI: uses speech service
README.md - FastAPI backend → Agent orchestration: submits utterance
app/main.py - Agent orchestration → Session context: uses context
app/agent/orchestrator.py - Agent orchestration → Voice guidance: uses guidance
app/agent/orchestrator.py - Agent orchestration → Safety guardrails: checks safety
app/agent/orchestrator.py - Agent orchestration → Tool registry: selects tools
app/agent/orchestrator.py - Agent orchestration → Session and events: records session
app/agent/orchestrator.py - Tool registry → Weather lookup: dispatches lookup
app/tools/registry.py - Tool registry → Government schemes: dispatches lookup
app/tools/registry.py - Tool registry → Courier tracking: dispatches tracking
app/tools/registry.py - Tool registry → Web search: dispatches search
app/tools/registry.py - Tool registry → SMS dispatch: dispatches SMS
app/tools/registry.py - Tool registry → Action lifecycle: normalizes actions
app/tools/registry.py - Agent orchestration → Source verification: verifies results
app/agent/orchestrator.py - Source verification → Official web sources: checks claims
README.md - Weather lookup → Open-Meteo: fetches forecast
README.md - Agent orchestration → AI model provider: requests reasoning
README.md - Agent orchestration → Speech synthesis: synthesizes answer
README.md - Speech synthesis → Sarvam AI: uses speech service
README.md - FastAPI backend → Document handoff: accepts upload
app/main.py - Document handoff → Image analysis: analyzes document
README.md - FastAPI backend → Source verification: verifies extracted claims
app/main.py - Web user → Live dashboard: uses interface
README.md - Live dashboard → FastAPI backend: uses backend
README.md - FastAPI backend → Session and events: persists session events
app/main.py - FastAPI backend → Phone caller: returns spoken audio
README.md