6 projects
pydantic-cal
LLM calibration metrics for pydantic-evals: ECE, MCE, ACE, Brier, reliability diagrams, temperature/Platt/isotonic scaling, semantic entropy.
f3dx
Rust runtime for Python AI apps. Drop-in for openai/anthropic SDKs with native SSE streaming, an agent loop with concurrent tool dispatch, and Logfire-compatible OTel emission.
f3dx-router
DEPRECATED. Consolidated into f3dx[router]. Install `f3dx` and import from `f3dx.router`.
f3dx-cache
DEPRECATED. Consolidated into f3dx[cache]. Install `f3dx` and import from `f3dx.cache`.
tracewright
Replay-driven eval for f3dx and pydantic-ai traces. Take a JSONL trace, swap the model, get a per-case diff.
llmkit-sdk
AI API gateway SDK with cost tracking, budget enforcement, and multi-provider routing for LLM applications
