8 projects
laidline-vllm
vLLM Model Runner V2 logits processor for the text watermark plus an ack-header middleware.
laidline-transformers
Laidline helper for Hugging Face model.generate: adds the text watermark logits processor and records text evidence.
laidline-pipecat
Pipecat `FrameProcessor` placed after TTS; marks TTS audio frames with streaming AudioSeal; one evidence entry per utterance.
laidline-livekit
LiveKit Agents `tts_node` wrapper: marks audio frames with streaming AudioSeal; one evidence entry per call.
laidline-diffusers
`attach(pipe)` marks every image a diffusers pipeline returns (watermark, C2PA, evidence) and replaces the SDXL invisible watermarker
laidline-comfyui
ComfyUI custom node `LaidlineSaveImage`: sign and watermark images on save (remote via LAIDLINE_URL or embedded via LAIDLINE_CONFIG).
laidline-bench
Robustness and calibration benchmarks for the Laidline watermarks; the code behind `laidline bench`.
laidline
Marking kit for generative AI providers: C2PA signing with your KMS, open watermarks, evidence log, free detect.
