ai-ml

Install local middleware proxies into Claude Code

Compress embeddings 8x instantly with TurboQuant. No training needed. Up to +8pp recall vs FAISS PQ at matched storage.

None

Converra SDK — transparent LLM client wrapping for conversation capture and A/B testing

Adversarial security testing framework for LLM-powered applications

Think-block filter for LLM streams

The missing friendly interface for BitNet inference. Ollama for 1-bit LLMs.

SpecLogician is an AI framework that turns code, tests, logs, and requirements into mathematical context for LLMs through formal specification synthesis, verification, and analysis.

The cLLM (chat-optimized Large Language Model, 'clem') framework tests such models' ability to engage in games, that is, rule-constituted activities played using language.

Production-ready Telegram FAQ bot with Russian LLMs, RAG, and multi-provider fallback

Portable Claude Code workspace archives

A Python package for tuning vLLM hyperparameters.

Self-hosted semantic search and knowledge management for LLM-driven development

Atomic Model Fragmentation (AMF) - Universal LLM Decomposition Library

Backend.AI Storage Proxy

LangChain BaseTool for estimating and deploying NEAR smart contracts

Canonical result and measurement data storage APIs for Cogniflow

LLMOps Observability SDK: decorators + SQS dispatch with compression

AI-powered Python fuzzer using LiteLLM and Atheris to automatically generate and execute fuzzing harnesses.

TurboQuant KV cache compression for local LLM inference