FAQ
Technical and product questions.
This side of Ellide is for people who care about document cleanup, structured outputs, and open-core workflows. The classroom-facing product lives on ellide.co.
Why use Ellide instead of uploading a PDF directly to an AI?
Raw PDFs are full of image data, noisy OCR, and layout junk. Ellide cleans that before the model sees it, which means more of the context window goes toward actual reasoning instead of recovery work.
What kinds of files does the hosted workflow target?
PDFs are the primary case, especially scans, reports, and slide exports. The broader product story also includes images, Word docs, and presentations, but the value prop here is simplest on OCR-heavy material.
Should I choose Markdown or JSON output?
Markdown is the best default for pasting into chat tools or storing readable cleaned text. JSON is better when your downstream system needs a more explicit structure.
Can I self-host or use my own API key?
That is the point of the open-core story. If you need to control credentials, inspect the code, or run the pipeline yourself, the GitHub repository is the right place to start.
How is data handled in the hosted workflow?
The product direction is that uploaded documents are processed for the task itself, not used to train models. Production privacy claims should continue to match the deployed backend exactly.
Who is this site for?
Business users, researchers, and developers who want a dead-simple way to clean documents for AI use. If you are looking for the educator-first tutoring workflow, head to the main site instead.