A benchmarking framework that evaluates the real effective context size of long-context LLMs using synthetic tasks that go beyond simple retrieval.

A unified playground for implementing and evaluating KV cache compression, retrieval, merging, and quantization methods to optimize long-context LLM inference memory usage.

A local code search engine that uses vector embeddings and ripgrep to enable semantic and keyword-based search across a codebase.

A local-first long-term memory runtime for macOS that creates an evidence-linked personal model of a user's work patterns and goals to provide grounded context for AI coding agents via MCP.

DLPack is an open in-memory tensor structure that acts as a common bridge for sharing tensors and operators across different deep learning frameworks.

A persistent memory layer for AI agents that enables shared context and recall across sessions, machines, and different agent runtimes.

An open-source platform for generating secure full-stack apps and AI agents from natural language, featuring a compiled runtime and built-in MCP server for self-hosted deployment.

A service quality monitoring tool for LLM APIs that verifies actual token generation to prevent "fake" uptime reports from proxy services.

An enterprise-grade AI Gateway framework powered by Azure API Management that provides centralized security, cost control, and observability for managing AI models, tools, and agents.

A lightweight API gateway that centralizes credential injection, rate limiting, and load balancing for services using API keys or bearer tokens.

An advanced Python deobfuscator that combines local LLMs with AST cleaning and pattern matching to help malware analysts and reverse engineers recover readable code.

A Kubernetes operator suite that automates the deployment, scaling, and management of LLM inference, fine-tuning, and RAG engines.

A distribution of manifests for installing the Kubeflow Platform on various Kubernetes clusters, providing a complete end-to-end machine learning environment.

A GPU-accelerated VLSI placement toolkit that treats chip component placement as a deep learning optimization problem to achieve massive speedups over traditional CPU tools.

An agent testing framework that uses simulations, user simulators, and judge agents to evaluate the behavior and robustness of AI agents in multi-turn conversations.

A modular runtime and orchestration system for AI agents that automates software development lifecycles through structured pipelines, quality gates, and specialized agent roles.

A set of tools for Mistral AI models that provides open-source tokenizers and Pydantic-based validation for text, images, and audio.

A benchmark for evaluating LLMs on real-world enterprise text-to-SQL workflows across multiple SQL dialects and complex data environments.

A high-performance constrained decoding library that enforces JSON schemas, regular expressions, and context-free grammars on LLM outputs to ensure structured data.

A terminal-based client that connects local Ollama LLMs to Model Context Protocol (MCP) servers, enabling local models to execute tools and access external resources.

An autonomous, AI-driven penetration testing framework that orchestrates over 72 security tools to automate reconnaissance, exploitation, and reporting.

A Model Context Protocol server that turns any Odoo 16+ database into an AI-ready interface, enabling LLMs to safely query and manage Odoo data without requiring Enterprise AI modules.

An experimental communications API for PyTorch that provides a high-level collectives interface and support for multiple hardware backends like NCCL, RCCL, and XCCL.

RelBench is an open benchmark for predictive machine learning on relational databases, providing a unified framework for loading and evaluating models on multi-table data.

A bridge that enables AI applications and agents to connect to AWS-hosted MCP servers by automatically handling AWS IAM (SigV4) authentication.

A high-performance JAX-based inference engine optimized for Google TPUs, providing high throughput and low latency for LLMs and multimodal models.

A framework for the symbolic distillation of deep neural networks, using PySR to approximate model components with symbolic equations for better interpretability.

A collection of portable, standardized skills for AI coding assistants that enable them to interact with databases, productivity tools, and cloud infrastructure.

An open-source data engineering harness that provides LLMs with deterministic tools for SQL analysis, column-level lineage, and warehouse connectivity across major cloud platforms.

A tool for tracking the energy consumption and greenhouse gas emissions of generative AI models used through APIs.

A sequence modeling toolkit for training custom content generation models, featuring a modular architecture and scalable training for models over 70B parameters.

A Swift library for running local LLMs on Apple platforms, providing high-level abstractions for structured output, tool use, and chain-of-thought reasoning.

An embeddable repository system that combines files, a SQL database, and version control into one ACID-compliant system for real-time collaboration and data history.

AMD MIGraphX is a graph inference engine that accelerates machine learning model inference on AMD hardware.

A collection of optimized pre-trained deep learning models and tools for accelerating the development and deployment of high-performance inference applications.

A tool for converting ONNX models into LiteRT, TFLite, TensorFlow, and native PyTorch code to facilitate cross-framework deployment.

A Python library for time series forecasting that integrates scikit-learn compatible models, statistical methods, and foundation models into a unified API.

An open-source, self-hosted agentic harness for internal AI teams to build sovereign AI workflows with sandboxed code execution and model independence.

Vearch is a cloud-native distributed vector database that enables efficient similarity search of embedding vectors for AI applications.

A ComfyUI node that dynamically adjusts reserved VRAM in real-time to prevent OOM errors and optimize GPU memory usage.

An OpenAI-compatible API proxy for the Tabbit browser, allowing users to access Tabbit's integrated AI models in third-party clients.

A containerized development environment that isolates Claude Code with bypassPermissions enabled, providing a secure sandbox for security audits and untrusted code exploration.

A privacy-focused collection of 152 browser-based utility tools, featuring on-device AI for transcription, translation, and image processing via Transformers.js.

A local proxy server for inspecting, replaying, and modifying network traffic from browsers, apps, and APIs, featuring an AI-powered control plane.

A collection of examples and tutorials for the ROCm platform, providing code for HIP runtime functionality, GPU-accelerated applications, and AI model execution.

An LLM-powered code completion package for Emacs that supports multiple providers and both chat-based and Fill-in-the-middle completion modes.

A Python CLI framework that stores threat‑hunting investigations as markdown using the LOCK pattern, makes them searchable, and optionally lets LLM agents generate research, run queries, and automate parts of the hunt.

A cost-effective inference library that enables running large Mixture-of-Experts models on memory-constrained GPUs by offloading expert weights to host memory and SSD.