imtt-dev/steer
The Active Reliability Layer for AI Agents. Catch failures, teach fixes, and automate reliability
Implements runtime verification through composable "judges" (JSON structure validation, PII detection, entropy-based slop filtering) that intercept model outputs and can be taught new rules via a local dashboard without code redeploys. Operates as a service mesh with framework-agnostic decorator patterns and zero-touch patching for PydanticAI, adding <5ms overhead while enforcing deterministic protocols locally. Captures failure-correction pairs for DPO training and generates audit logs with deterministic reason codes for every blocked response.
130 stars.
Stars
130
Forks
3
Language
Python
License
MIT
Category
Last pushed
Jan 19, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/imtt-dev/steer"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
petterjuan/agentic-reliability-framework
ARF is an agentic reliability intelligence platform that separates decision intelligence (OSS)...
sarkar-ai-taken/riva
Local-first observability and control plane for AI agents.
Nubaeon/empirica
Make AI agents and AI workflows measurably reliable. Epistemic measurement, Noetic RAG,...
relai-ai/relai-sdk
A platform for building reliable AI agents
soumendrak/ragwatch
An SDK for Python AI Agents. Under heavy development.