The Windows GPU server has been the actual production home for these
services for some time, and the running code there has drifted ahead of
the repo. This sync pulls the live versions back into the repo so the
Windows box is no longer the only place those changes exist.
Pulled from C:\mana\services\* on mana-server-gpu (192.168.178.11):
mana-llm:
- src/main.py, src/config.py — small fixes (auth wiring, config tweaks)
- src/api_auth.py — NEW (cross-service GPU_API_KEY validator)
- service.pyw — Windows runner used by the ManaLLM scheduled task
(sets up logging redirect, loads .env, calls uvicorn)
mana-stt:
- app/main.py — substantial cleanup (684→392 lines), drops the
whisperx-as-separate-backend branching now that whisper_service.py
rolls whisperx in directly
- app/whisper_service.py — full CUDA + whisperx rewrite (158→358 lines)
- app/auth.py + external_auth.py — significantly expanded auth
- app/vram_manager.py — NEW (shared VRAM accounting helper)
- service.pyw — Windows runner with CUDA pre-init, FFmpeg PATH
injection, .env loading
- removed: app/whisper_service_cuda.py (folded into whisper_service.py)
- removed: app/whisperx_service.py (folded into whisper_service.py)
mana-tts:
- app/auth.py, external_auth.py — same auth expansion as stt
- app/f5_service.py, kokoro_service.py — Windows tweaks
- app/vram_manager.py — NEW (same shared helper as stt)
- service.pyw — Windows runner
mana-video-gen:
- service.pyw — Windows runner (no other changes; the .py code on the
GPU box is byte-identical to what's already in the repo)
The service.pyw files contain absolute Windows paths
(C:\mana\services\<svc>) and a hardcoded FFmpeg PATH for the tills user
profile. Kept as-is intentionally — they exist to be deployed to that
one machine and any abstraction layer would just hide what's actually
happening. Anyone redeploying to a different layout will need to edit
the path strings, which is a known and obvious change.
Mac-Mini infrastructure for these services (launchd plists, install
scripts, scripts/mac-mini/setup-{stt,tts}.sh, the Mac-flux2c image-gen
implementation) is still on disk and will be removed in a follow-up
commit, along with replacing mana-image-gen with the Windows
diffusers+CUDA implementation. This commit is just the live-code sync.
Create packages/shared-python/manacore_auth/ with:
- auth.py: API key validation, rate limiting, local + external auth
- external_auth.py: mana-core-auth remote validation with caching
- create_auth_dependency(scope): factory for per-service auth deps
Migrated services:
- mana-stt: auth.py now wraps shared auth with scope="stt" (272→42 LOC)
- mana-tts: auth.py now wraps shared auth with scope="tts" (272→42 LOC)
The only difference between services was the scope parameter ("stt" vs "tts").
Both external_auth.py files were 100% identical and are now thin re-exports.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
mana-stt: add WhisperX service with CUDA GPU support, speaker diarization, and auto-fallback chain.
mana-notify: add locale fallback and default templates for task reminders.
CD: update deployment pipeline and docker-compose configuration.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add api_keys schema in mana-core-auth with SHA-256 hashing
- Create NestJS module with CRUD endpoints and validation
- Add external auth module to STT/TTS for sk_live_ key validation
- Create web UI page at /api-keys for key management
- Support rate limiting per key with configurable limits
- Cache validation results for 5 minutes to reduce auth service load
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add auth.py module to both STT and TTS services with:
- API key validation via X-API-Key header
- Rate limiting with sliding window (requests per minute)
- Internal API key option for unlimited access
- Environment variable configuration
All protected endpoints now require authentication.
Public endpoints (/health, /docs) remain accessible.
- Disable vLLM by default (has issues on macOS CPU)
- Use Mistral API for Voxtral transcription (cloud-based)
- Keep Whisper-MLX for local transcription
- Update README with architecture diagram
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add vllm_service.py as proxy to vLLM server for Voxtral 3B/4B
- Add voxtral_api_service.py for Mistral API fallback
- Update main.py with /transcribe/voxtral endpoint using vLLM
- Add /transcribe/auto endpoint with automatic fallback chain
- Create setup-vllm.sh and start-vllm-voxtral.sh scripts
- Add launchd plist files for Mac Mini deployment
- Add install-services.sh for automated service installation
Architecture:
- vLLM server runs Voxtral models on port 8100
- mana-stt proxies to vLLM with Mistral API fallback
- Fallback chain: vLLM -> Mistral API
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add mana-stt service with Whisper and Voxtral support for local
transcription. Includes setup script and launchd integration for
automatic startup on Mac Mini server.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
