SnapOtter

SnapOtter is a self-hosted application for people and teams who need to edit or convert private files without uploading them to an online processing service. It combines media, document and local AI tools in a browser interface.

• Media editing: Resize and compress images, trim and merge video, or clean up audio; image formats include camera RAW, HEIC, WebP and AVIF.
• Documents and files: Merge, redact, sign and repair PDFs, convert Office documents, strip metadata, and work with CSV, JSON, XML, YAML and ZIP files.
• Image editor: Edit images with layers, brushes, shapes, curves and filters while processing stays on your hardware.
• Offline AI: Remove backgrounds and objects, upscale or restore photos, extract text from images and PDFs, transcribe audio, and generate video subtitles without external AI APIs.
• Automation: Connect tasks into reusable pipelines, process files in batches, or integrate the tools with internal software through a REST API and completion webhooks.

It runs through Docker, Kubernetes, bare metal or a cloud VM you control, with ARM and x86 support on Linux and macOS. Processing can operate fully offline, including in air-gapped environments. Files remain on your infrastructure.

AI tools can run on CPU. NVIDIA CUDA accelerates background removal, upscaling and transcription; OCR uses CPU, and Intel or AMD integrated GPUs don't accelerate AI inference. The project is free and open source under AGPL-3.0, with commercial licensing also available.