Parse
Tables stay tables.
Layout-aware Markdown with source locations.

Perception series
Five tools that parse, split, classify, extract and ground any of 36 file types, and point every value back to the page it came from.
Ranks first on grounding against frontier models. Available in the Console today.


Page 1 of 1 · 0 regions
Our most accurate setting reads documents as well as the best tool we tested, for less than half the price. Our lightest setting is the cheapest of them all and still beats all but one of the leading AI models.
Parsing leaderboard export, 15 September 2026. Cost is USD per 1,000 pages as measured. ParseBench is scored on its main metric, Content Faithfulness, only.
Tables stay tables.
Layout-aware Markdown with source locations.

One packet becomes the documents inside it.
Page ranges for each document in a mixed upload.
Labels come from your taxonomy.
Document or page labels from your taxonomy.
Every field says found, not found or ambiguous.
JSON Schema output, a status and a citation on every field.

A location, not a judgment.
Any quote resolved to its page, cells or span.


PDF, Word, PowerPoint, images and scans, Visio

Excel, CSV, XML, read cell by cell

Text, HTML, email, audio and video
Same models, same API at every tier. Only the boundary around your data moves.
01
Our GPUs, our API, a key from Console.
02
Your own cluster in our cloud, reached over private links.
03
Installs in your AWS or Azure account. Data stays inside.
04
Your datacenter, your hardware, zero outbound.
