Purpose and audience
This browser agent turns visible page controls into an indexed action space for a natural-language goal.
Evidence 1 ↗ · Evidence 2 ↗
How Jev fits
Jev chooses the operation and compatible element in one request; a separate small LLM generates text for TYPE_TEXT.
Evidence 1 ↗ · Evidence 2 ↗
Using the project
Use the local inspector or Python Agent interface, with Chrome connected through Browser Harness and separate decision and text-model credentials.
Evidence 1 ↗ · Evidence 2 ↗
Scope and limits
DONE still needs outcome verification. Frames, shadow roots, uploads and pop-up tabs are outside the MVP; reported timings cover specific tasks.
Evidence 1 ↗
Evidence checked:
Project facts
- Relationship to Jev
- Jev application
- Evidence
- Documented
- Language
- Python
- License
- MIT
- Origin
- Original repository
- Repository status
- Not archived
- Created
- 2026-09-16
- Last code update
- 2026-09-18
- GitHub stars
- 21,030
- Evidence checked
- Sep 30, 2026
- Metadata checked
- Sep 28, 2026
- Check status
- Up to date
Stars measure the whole repository, including work unrelated to Jev.
Evidence and scope
Documented records the linked documentation or source. JevHunt has not independently run or benchmarked this project.
Evidence commit: 1231850a0bf1a0c0341fe408ef1668dbbfdfac46
Discovered through: awesome-jev, typesafe-field-guide.
How we review →