Skip to content

KNOWN_LIMITATIONS.md

Latest commit

 

History

History
50 lines (34 loc) · 3.16 KB

KNOWN_LIMITATIONS.md

File metadata and controls

50 lines (34 loc) · 3.16 KB
Raw

Known Limitations

Missing owner-only surfaces

The public edition does not include the commercial product's owner-only features:

  • Discord Clip Room bot and delivery (the Discord status UI, "Send to Discord" action, /discord/* API client surface, and api.clipRoom.deliver method are not present in this edition). The public Clip Room also omits the backend POST /clip-room/candidates/{id}/send route and the corresponding send action in the frontend api.clipRoom.action surface; the underlying send_to_discord state-machine transition remains as a local no-op through NullDiscordGateway for source parity.
  • Twitch sync and credentials (the /twitch/* API client surface, including status, credentials, and sync methods, is not present in this edition).
  • Publishing to Instagram, TikTok, or YouTube (the publishing UI, account management, queue, and /publish/* API client surface are not present in this edition).
  • Cloud tester gateway (the LLM judge is opt-in via CLIPPER_GATEWAY_URL only).
  • Creator OS surfaces: Director chat, Memory UI, Stream Studio, Research Desk, Idea Bank.
  • Release packaging, code signing, and installer machinery.

Verification gaps

The following were not run during preparation of this public edition:

  • tauri build (Rust desktop bundle).
  • Windows installer / NSIS / MSI creation.
  • Clean-machine acceptance testing on a fresh OS install.
  • Live cloud gateway tests (no CLIPPER_GATEWAY_URL was configured).
  • End-to-end processing of a real multi-hour VOD in this verification pass.

Dependency constraints

  • mediapipe requires Python 3.12 at the time of writing and has no published wheels for Python 3.13/3.14.
  • The verification environment used Python 3.14.6.
  • The following heavy ML/audio packages were not installed there: torch, mediapipe, resemblyzer, sounddevice, opencv-python.
  • The test suite passes without them because modules degrade gracefully and tests mock the heavy paths.
  • For full runtime features (face detection, speaker ID, quality classifier training) install Python 3.12 and the packages listed in requirements.txt.

Runtime behavior

  • The faster-whisper model downloads from HuggingFace on first use.
  • Face/scene models auto-download from public sources when their dependencies are installed.
  • The sound FX library is empty in this edition; no redistributable audio assets are included.
  • Tauri icons are not committed. Run frontend/generate_icons.py from a source icon before tauri build.

Security / dependency audit

  • npm audit reports 4 vulnerabilities in upstream dependencies (1 low, 3 high) as of the verification run. These were not introduced by this edition's code.
  • See docs/PRIVACY_AND_SECURITY.md for the local-API guard details.

Edition-specific exclusions

  • Cloudflare quick-tunnel file sharing (modules/tunnel_publish.py) is not included in this edition; when a rendered clip exceeds the attachment limit, delivery logs a local notice instead of exposing a public URL.

License

No license has been chosen yet; all rights reserved pending an owner decision. Do not use, distribute, or create derivative works without explicit permission from the owner.