Products
Platforms built for production, not demos.
Every MenteE platform is engineered for reliability at scale — designed to be used, not just shown.
MenteE Embed
Compact multilingual text embedding model — ~41M params, trained from scratch for Arabic, English, and Urdu retrieval. 384-dimensional outputs, up to 512-token context, fully reproducible. Apache 2.0.
~41M
Parameters
384
Dimensions
3
Languages
Apache 2.0
License
MenteE SWE
BetaAutonomous software-engineering agent that lives in your terminal. Investigates your repository, plans changes, edits files, verifies with your tests, and reports with evidence. Model-agnostic, safe by design.
Model-agnostic
Kimi, GLM, Z.ai
Safe
No silent deletions
Persistent
Memory across sessions
MIT
Licensed