Potato is an open-source platform for annotating text, audio, images, video and qualitative data. Annotation tasks are set up in YAML, and its maker says an annotation interface can be built without programming. The homepage lists 61 annotation types, including classification, spans, bounding boxes, polygons, video, audio and qualitative coding. Potato supports annotation of multi-agent teams on an interaction graph and evaluation of computer-use, voice and video agents. AI assistance includes label suggestions, keyword highlighting, quality checks, pre-annotation, explanation generation and consistency checks. Listed integrations include OpenAI, Anthropic Claude, Google Gemini, Ollama, HuggingFace, OpenRouter, vLLM, YOLO and LangChain. Imports cover text, images, audio, video and PDF or HTML documents; exports include JSON, JSONL, CSV, CoNLL, Hugging Face, COCO and YOLO. Potato runs locally or on users’ own servers, which its maker says keeps data on their infrastructure and supports offline use. The free self-hosted plan includes all features with no usage limits. It is released under GNU GPL version 3 or later; installation with pip requires Python 3.7 or later.
Who it is for
Potato suits teams that need to annotate different media types or evaluate AI agents and can run software on their own infrastructure. It may also fit projects requiring offline use or control over where data is kept.
What is good
- Lists 61 annotation types across multiple media.
- AI assistance includes suggestions, pre-annotation and consistency checks.
- Imports and exports a range of data formats.
- Runs locally or on users’ own servers, with offline use supported.
- Free plan has all features and no usage limits.
What to know first
- Self-hosted only; users must run it themselves.
- No managed cloud tier.
- Installation requires Python 3.7 or later.
Verdict
Potato offers broad annotation and AI-assisted labeling options in a free, self-hosted package. It is a fit for users prepared to run and maintain it themselves; there is no managed cloud tier.
Potato plans and pricing
All plansCompared on data labeling software
- Free plan
- Yes
- Image annotation
- Yes
- Text annotation
- Yes
- Audio/video annotation
- Yes
- Model-assisted labeling
- Yes
- Review workflow
- Yes
- API or SDK access
- Yes
- Deployment
- self-hosted
