ArchishmanSengupta/autovoiceevals

A self-improving loop for voice AI agents. Uses karpathy's autoresearch as foundation.

47
/ 100
Emerging

Uses Claude to generate adversarial caller scenarios and iteratively refine agent prompts—proposing one change at a time, executing the modified prompt against all test cases, and keeping/reverting based on score improvement. Integrates directly with Vapi, Smallest AI, and ElevenLabs ConvAI via their APIs, automatically reading and updating live agent prompts while preserving the original on completion.

No Package No Dependents
Maintenance 13 / 25
Adoption 9 / 25
Maturity 9 / 25
Community 16 / 25

How are scores calculated?

Stars

83

Forks

13

Language

Python

License

MIT

Last pushed

Mar 18, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/ArchishmanSengupta/autovoiceevals"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.