Popular repositories Loading
-
evalharness
evalharness PublicA small, reusable LLM evaluation harness — golden test sets, LLM-as-judge scoring, regression detection against a baseline, and a CI quality gate that fails the build when quality drops.
Python
-
aegisquery-mcp
aegisquery-mcp PublicGoverned analytics MCP server: safe read-only LLM access to a data warehouse with a layered safety pipeline (AST validation, PII masking, cost limits, audit) over stdio and OAuth HTTP.
TypeScript
-
Dialoft
Dialoft PublicAI Voice SDR — one LangGraph brain behind an OpenAI-compatible API drives both phone (Vapi) and browser (LiveKit): BANT qualification, objection handling, meeting booking, and CRM logging, with con…
Python
-
penumbra
penumbra PublicProbabilistic space-weather driver forecasting - F10.7 and Kp forecasts with calibrated uncertainty bands, translated into satellite orbital-drag risk. Phase-2 companion to KESSLER.
Python
-
anvaya
anvaya PublicAI failure observability - detect, classify and attribute failures in AI products. 57 research-grounded failure modes, tiered detection, causal attribution.
TypeScript
If the problem persists, check the GitHub status page or contact support.