0 Fork 0

mirror of https://github.com/MervinPraison/PraisonAIUI.git synced 2026-11-13 18:48:37 -10:00

Files

T

MervinPraisonandClaude Opus 4.8 3608d011c6 fix(voice-agent): build stable webhook idempotency keys so status updates aren't dropped

webhook_event_key derived the dedup key from a 200-char JSON prefix for
status-update/conversation-update/end-of-call-report. When a large static
field (e.g. `call`) sorts before the discriminating field, successive
events collapse to the same prefix and later ones are dropped as
duplicates, freezing the dashboard status on the first value (e.g. stuck
on 'ringing'). Derive keys from the fields that actually distinguish each
event, matching the existing transcript-event pattern. No new deps,
params, or exports.

Co-Authored-By: Claude Opus 4.8 <[email protected]>

2026-09-07 23:45:23 +00:00

integrations

fix(voice-agent): build stable webhook idempotency keys so status updates aren't dropped

2026-09-07 23:45:23 +00:00

.env.example

feat(voice-agent): Phase 1 telephony webhooks, tools, and call dashboard

2026-09-07 13:18:37 +05:30

app.py

fix(voice-agent): remove committed run logs, timing-safe webhook verify, lazy httpx

2026-09-07 10:06:59 +00:00

README.md

feat(voice-agent): Phase 1 telephony webhooks, tools, and call dashboard

2026-09-07 13:18:37 +05:30

setup_all.ps1

feat(voice-agent): Phase 1 telephony webhooks, tools, and call dashboard

2026-09-07 13:18:37 +05:30

setup_voice_provider.py

feat(voice-agent): Phase 1 telephony webhooks, tools, and call dashboard

2026-09-07 13:18:37 +05:30

start_dev.ps1

feat(voice-agent): Phase 1 telephony webhooks, tools, and call dashboard

2026-09-07 13:18:37 +05:30

voice_routes.py

fix(voice-agent): hold strong ref to background webhook task to prevent GC drop

2026-09-07 10:16:51 +00:00

README.md

Voice Agent example — PraisonAIUI

Phone and web voice agents using an external telephony/voice platform for STT/TTS/calls and PraisonAI for tools + dashboard.

Quick start

cd examples/python/voice-agent
copy .env.example .env
# Fill VOICE_API_BASE, VOICE_PRIVATE_API_KEY, VOICE_ASSISTANT_ID, etc.

# Fully automatic (app + tunnel + provider config):
.\setup_all.ps1

# Or step by step:
.\start_dev.ps1

Dashboard: http://127.0.0.1:8001

Provider setup

Create an assistant with tools get_current_time and echo_message
Set webhook URL to {PUBLIC_API_BASE_URL}/webhooks/voice
Set webhook secret header X-Voice-Webhook-Secret → same as VOICE_SERVER_URL_SECRET in .env
Add a phone number and copy VOICE_PHONE_NUMBER_ID

Or run py -3.13 setup_voice_provider.py after start_dev.ps1.

API

Route	Method	Purpose
`/api/voice/calls`	POST	Outbound call `{ "customer_number": "+1..." }`
`/api/voice/calls`	GET	List stored calls
`/api/voice/calls/{id}`	GET	Call detail + transcript
`/webhooks/voice`	POST	Provider webhooks (tool-calls, transcript, end-of-call)
`/api/voice/config`	GET	Config status

Architecture

External platform handles voice pipeline (phone, STT, TTS, turn-taking). This app handles:

tool-calls → PraisonAI-registered tools (integrations/voice/tools.py)
transcript / end-of-call-report → SQLite store + dashboard pages
assistant-request → returns VOICE_ASSISTANT_ID

Same pattern as examples/python/meeting-agent (Recall webhooks).

Tests

pytest tests/unit/test_voice_agent.py -v