Zipstack/unstract
A self-hosted platform that turns PDFs, scans and spreadsheets into structured JSON. You describe the fields in plain language in its Prompt Studio, then call an API or run batch jobs over a folder. It works with several LLM providers and vector databases.
Not tested yetREADME only
How we know
We only read the README and a summary and tested nothing, so there is no verdict yet.
Good for
Teams with piles of invoices, forms or contracts who want the data kept on their own servers.
Not for
A handful of one-off files. It runs a full stack of Postgres, Redis, RabbitMQ and Celery, and the AGPL-3.0 license needs a check before you ship it inside a product.
Repo
- Stars
- 7,230
- License
- AGPL-3.0Like GPL, and running it as a web service counts too: people who use your service can ask for your source code.
- Maturity
- active
- Last push
- 2026-09-10
- Last checked
- 2026-11-13