New · under observation
datalab-to/indra
1M param OCR model - runs on CPU, 15x faster than tesseract, close to 1B+ param models in accuracy
- Repository created
- Oct 13, 2026
Observing since Nov 7, 2026. Final review from Dec 7. Listed automatically.
Activity
Daily measurements 2 readings
How measurements work
Hourly checks may be delayed by hosting or rate limits. Times are UTC. Daily views include today in UTC. Daily totals use the last reading; commits sum intervals ending that day. Missing readings are not zero activity.
| Date · UTC | Stars | Star change | Forks | Open issues & PRs | Commits / interval |
|---|---|---|---|---|---|
| Nov 8, 2026 20:10 | 1,682 | +17 | 173 | 9 | 0 / 20.9h |
| Nov 7, 2026 23:13 | 1,665 | — | 170 | 9 | 0 / 3.1h |
Plan a pilot
Download a brief to share with the team. Nothing is sent or saved.
Developments
Past 30 days · Checked Nov 8, 20:22 UTC
- API submits and result checks, 2x faster ↗
- Which document-parsing benchmarks can you trust? ↗
- Data extraction vs parsing: Which to use? ↗
- Surya 2.5: 1.7x faster, best open OCR under 1B ↗
- Chandra 2.6: more accurate than the frontier, 2.5x faster, same price ↗
- Lite vs Azure, AWS and Google: 2x faster, 10x cheaper, as accurate ↗
- Open-source OCR models on OmniParseBench ↗
- Lite mode: document parsing for $0.50 per 1,000 pages ↗
- Improved document segmentation ↗
- Word boxes that hold up on handwriting ↗
Reported by linked sources.
About these listings
Workflow categories are inferred from repository descriptions. New listings are under observation and await a final review after 30 days. Product capabilities, funding and customer adoption are unverified.
GitHub activity reflects development and attention. It does not establish product quality or business results.