Research + corpus-building uses. Benchmarks don’t count.
The drawers
A tool can be filed in more than one drawer
DRAWER I
AI & History
Machine learning, vision-language models and speech models. Every card here carries notes on where your sources go, how it was tested, and how it fails.
Guide cards
Counts research and corpus-building uses in the window, plus “I’m using this now” check-ins. Benchmarks don’t count.
tool · No. 017AI
olmOCR
Fully open toolkit from the Allen Institute for AI that turns PDFs and scans into clean, reading-order text, built to process millions of pages cheaply.
At a glance
Maker
Allen Institute for AI
Released
Feb 2025 (v2: Oct 2025)
Licence
Open source (Apache-2.0)
Cost
Free; third-party hosting available
Runs
Your own GPU (12 GB+); hosted APIs; free web demo
Skill
Some code
Historian’s notes
Where your sources goAi2’s terms give it broad rights over content submitted to its services. Run it locally for private sources.
How it’s been tested1.93% error on clean 19th-century print, 4.27% on early-modern print (Clifford et al. 2026).
Known problems on historical sourcesSilently modernises old spelling, recovers only about half the text on multi-column pages, and invents names (“Gotland” became “Goliath”). English-focused.
Used in research0
No published research uses found yet. Using it in a project? Tell us.
Nobody has said they’re using it in 2026 yet.
“I’m using this now” asks you to sign in with ORCID.
One count per person per year; it adds to Trending, and names are never shown.