Readiness check

Readiness Check

Verify that all OpenCode plugin services are healthy and operational.

Quick Health Check

Run these commands to verify all services:

# 1. Check Whisper STT server
curl -s http://localhost:8787/health

# 2. Check Coqui TTS server
echo '{"text":"test", "output":"/tmp/test.wav"}' | nc -U ~/.config/opencode/coqui/tts.sock

# 3. Check running processes
ps aux | grep -E "whisper_server|tts_server" | grep -v grep

# 4. Check Supabase RLS (requires .env with SUPABASE_ANON_KEY)
source .env && curl -s "https://slqxwymujuoipyiqscrl.supabase.co/rest/v1/telegram_replies?select=id&limit=1" \
  -H "apikey: $SUPABASE_ANON_KEY" -H "Authorization: Bearer $SUPABASE_ANON_KEY"

# 5. Check Supabase migrations are in sync
supabase migration list

Service Details

Whisper STT Server

Purpose: Transcribes voice messages from Telegram to text.

Location: whisper/whisper_server.py

Default Port: 8787

Start Command:

cd /path/to/opencode-reflection-plugin/whisper
python3 whisper_server.py --port 8787 &

Health Check:

curl -s http://localhost:8787/health

Expected Response:

{
  "status": "healthy",
  "model_loaded": true,
  "current_model": "base",
  "available_models": ["tiny", "tiny.en", "base", "base.en", "small", "small.en", "medium", "medium.en", "large-v2", "large-v3"]
}

Troubleshooting:

• If not running: Start with the command above
• If model loading fails: Check Python dependencies (pip install openai-whisper)
• For faster startup: Use --model tiny (lower quality but faster)

---

Coqui TTS Server

Purpose: Generates speech audio from text responses.

Location: ~/.config/opencode/coqui/tts_server.py

Socket Path: ~/.config/opencode/coqui/tts.sock

PID File: ~/.config/opencode/coqui/server.pid

Health Check:

# Check socket exists
ls -la ~/.config/opencode/coqui/tts.sock

# Check process is running
cat ~/.config/opencode/coqui/server.pid
ps aux | grep "$(cat ~/.config/opencode/coqui/server.pid)"

# Test TTS generation
echo '{"text":"Hello, this is a test.", "output":"/tmp/test_tts.wav"}' | nc -U ~/.config/opencode/coqui/tts.sock

Expected Response:

{"success": true, "output": "/tmp/test_tts.wav"}

Verify Audio:

# Check file was created
file /tmp/test_tts.wav
# Expected: RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, mono 48000 Hz

# Play audio (macOS)
afplay /tmp/test_tts.wav

Troubleshooting:

• If socket missing: The TTS plugin auto-starts the server on first use
• To manually restart: kill $(cat ~/.config/opencode/coqui/server.pid) then trigger TTS
• Check logs in ~/.config/opencode/coqui/

---

Plugin Deployment

Plugin Location: ~/.config/opencode/plugin/

Check Deployed Plugins:

ls -la ~/.config/opencode/plugin/

Expected Files:

• reflection-3.ts - Judge layer for task verification
• tts.ts - Text-to-speech with Telegram integration

Deploy from Source:

cp /path/to/opencode-reflection-plugin/tts.ts ~/.config/opencode/plugin/
cp /path/to/opencode-reflection-plugin/reflection-3.ts ~/.config/opencode/plugin/reflection.ts

Restart OpenCode after deploying for changes to take effect.

---

TTS Configuration

Config File: ~/.config/opencode/tts.json

View Current Config:

cat ~/.config/opencode/tts.json

Example Configuration:

{
  "enabled": true,
  "engine": "coqui",
  "os": {
    "voice": "Samantha",
    "rate": 200
  },
  "coqui": {
    "model": "jenny",
    "device": "cpu",
    "language": "en",
    "serverMode": true
  },
  "telegram": {
    "enabled": true,
    "uuid": "your-uuid-here",
    "sendText": true,
    "sendVoice": true,
    "receiveReplies": true
  },
  "whisper": {
    "enabled": true,
    "model": "base",
    "port": 8787
  }
}

---

Supabase Backend Verification

RLS Policy Check

The telegram_replies table requires proper RLS policies for:

• SELECT with anon key (enables Realtime subscriptions)
• mark_reply_processed RPC function (marks replies as handled)

Test SELECT Policy:

source .env && curl -s "https://slqxwymujuoipyiqscrl.supabase.co/rest/v1/telegram_replies?select=id,uuid,processed,created_at&order=created_at.desc&limit=3" \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" | jq '.'

Expected: Array of reply objects (not an error)

Test RPC Function:

source .env && curl -s "https://slqxwymujuoipyiqscrl.supabase.co/rest/v1/rpc/mark_reply_processed" \
  -X POST \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" \
  -H "Content-Type: application/json" \
  -d '{"p_reply_id": "00000000-0000-0000-0000-000000000000"}' | jq '.'

Expected: true or false (not a permission error)

Migration Sync Check

supabase migration list

Expected: All migrations show both Local and Remote columns with matching timestamps.

If migrations are out of sync:

# If remote has migrations not in local
supabase migration repair --status reverted <migration_id>

# Then push local migrations
supabase db push

Edge Functions Check

# List deployed functions
supabase functions list

# Check function logs
supabase functions logs telegram-webhook --tail
supabase functions logs send-notify --tail

---

Telegram Integration Verification

1. Check Telegram Config

cat ~/.config/opencode/tts.json | jq '.telegram'

Required fields:

• enabled: true
• uuid: Your user UUID from /start command
• receiveReplies: true (for two-way communication)

2. Test Outbound Notifications

Trigger a TTS event and check if notification was sent:

# Check recent notifications in Supabase
source .env && curl -s "https://slqxwymujuoipyiqscrl.supabase.co/rest/v1/telegram_notifications?select=id,message,created_at&order=created_at.desc&limit=3" \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" | jq '.'

3. Test Inbound Replies

Send a message to the Telegram bot, then check if it appears:

# Check for unprocessed replies
source .env && curl -s "https://slqxwymujuoipyiqscrl.supabase.co/rest/v1/telegram_replies?select=id,reply_text,is_voice,processed,created_at&processed=eq.false&order=created_at.desc&limit=5" \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" | jq '.'

4. Test Voice Transcription

Send a voice message to the bot, then verify transcription:

# Check if Whisper is running
curl -s http://localhost:8787/health | jq '.'

# Voice messages will have is_voice=true and reply_text populated after transcription

---

Common Issues

Issue	Cause	Solution
Whisper not responding	Server not started	python3 whisper_server.py --port 8787 &
Coqui socket missing	Server not started	Trigger any TTS action or restart OpenCode
Supabase module error	Dependency missing	Add to ~/.config/opencode/package.json and run bun install
Telegram not working	Missing UUID	Get UUID from Telegram bot with /start
Voice messages not transcribed	Whisper not running	Start Whisper server
RLS permission denied	Missing SELECT policy	Deploy 20240117000000_fix_replies_rls.sql migration
Realtime not receiving	Anon key blocked by RLS	Deploy RLS fix migration with SELECT policy for anon
mark_reply_processed fails	RPC function missing	Deploy RLS fix migration with SECURITY DEFINER function
Migrations out of sync	Remote has unknown migrations	Run supabase migration repair then supabase db push