An AI-native, cloud-native distributed file system that provides a high-performance POSIX layer and multi-tier caching over cloud object storage for AI training and Agent platforms.

The Python SDK for the Gemini Enterprise Agent Platform (formerly Vertex AI) allows developers to build, deploy, and optimize AI agents, manage prompts, and evaluate generative AI models.

A suite of test scenarios and games for multi-agent reinforcement learning to evaluate how AI agents generalize social interactions to novel situations.

An enterprise-grade secure gateway that provides auditing, governance, and access control for AI API calls and MCP tool invocations across an organization.

An agentic AI security framework for deep security code review that uses LLMs to detect subtle vulnerabilities and reduce false positives through integrated triage.

A JAX-native library for Multi-Agent Reinforcement Learning that provides GPU-accelerated environments and baseline algorithms for efficient benchmarking.

A native Swift macOS menu bar app for tracking Claude Code token consumption, costs, and server-side usage limits in real-time.

A curated collection of pre-compiled Python wheels for AI/ML libraries that are difficult to install on Windows, eliminating the need to build from source.

Deltafin is a Rust‑based native binary that runs the full 2.8 T‑parameter Kimi K3 Mixture‑of‑Experts model on consumer hardware (macOS arm64 or Linux x86‑64/aarch64). It keeps all 16 experts, offers both full‑download and on‑demand streaming installations, optional Qwen draft‑model acceleration, a CLI for chat/completion, and an OpenAI‑compatible server. The project is MIT‑licensed, with model weights under their original terms.

A Python package for approximating any-order Shapley interactions to explain feature synergies in machine learning model predictions.

A CCIE-level AI network engineering agent that automates monitoring, troubleshooting, and configuration across multivendor networks using hundreds of skills and MCP integrations.

A local LLM routing proxy that reduces API costs by 40-70% by routing prompts to the cheapest viable model and escalating to premium models only when necessary.

A GPU library built on Triton and CuteDSL that provides high-performance implementations of classical machine learning operators like k-means, PCA, and ANN search.

A minimal and scalable JAX-based research codebase designed for rapid LLM iteration and automated AI research conducted by both humans and AI agents.

A high-performance AI operator library for AMD ROCm that provides optimized GPU kernels for AI inference and training workloads.

A utility package for the Apple Foundation Models framework that provides chat completions clients, context window management, and dynamic skill activation for LLMs.

A React-based UI kit providing components and templates for quickly building and deploying user interfaces for Pipecat AI voice applications.

A pytest-native safety and security testing framework designed for red teaming and risk assessment of agentic AI applications.

A local-first desktop application for running and comparing LLM performance across various benchmark sets called Bench Packs against local or remote models.

A Clang compiler plugin for C++ that provides automatic differentiation by transforming source code to generate functions for derivatives, gradients, and Hessians.

A GPU-accelerated library for single-cell analysis that integrates with the scverse ecosystem and uses NVIDIA RAPIDS to speed up computations on large datasets.

A port of Andrej Karpathy's llm.c to the Mojo language, demonstrating that Mojo can achieve C-like performance for low-level LLM training.

A visual debugger for Qwen3.6-27B on Apple Silicon that uses a Jacobian lens to visualize and edit the model's internal latent representations.

A TUI-based model manager and server launcher that auto-detects inference backends and local model files to simplify serving LLMs.

A lightweight background daemon that connects local AI agents to Telegram and WeChat, allowing users to to monitor and control agents remotely via mobile notifications.

An LLM evaluation framework featuring benchmark environments, a pluggable interceptor proxy for traffic management, and multi-format reporting for model performance analysis.

FlagTree is a unified open-source compiler for multiple AI chips that enables a "develop once, run anywhere" workflow by unifying various Triton-based backends.

A utility library for Unsloth that enables faster and more memory-efficient fine-tuning of LLMs, vision, and TTS models with significantly reduced VRAM requirements.

A local macOS server that exposes Cursor's first-party models, such as Composer 2.5 and Grok 4.6, as OpenAI-compatible API endpoints.

A network observability tool for Kubernetes that uses eBPF to index traffic and provides AI-driven troubleshooting via MCP integration.

A professional context and harness engineering system that wraps around Claude Code and Codex to provide durable context, quality gates, and runtime verification for production-grade software.

An open-source agentic code reviewer that performs deep architectural audits using a dynamic pipeline and evidence grounding to find systemic vulnerabilities.

A GPU-accelerated library for vector similarity search and clustering, providing high-performance implementations of approximate nearest neighbors algorithms for AI and data mining.

An enterprise search server built on OpenSearch that crawls and indexes documents from websites, file systems, and cloud data stores for centralized full-text search.

A structural memory system for AI coding agents that turns codebases into local, bi-temporal knowledge graphs for fast, low-cost architectural queries.

A high-performance Julia framework for agent-based modeling that enables the simulation of autonomous agents reacting to their environment and each other.

A comprehensive collection of 20+ tools, function pipes, and filters for Open WebUI that adds academic research, multi-modal generation, and autonomous agent capabilities.

A markdown filesystem that makes documentation writable, searchable, and versioned for both AI agents and humans, featuring native MCP support and Git-based auditing.

An inference-native agent harness designed for loop engineering, optimizing token usage and cache reuse to maintain efficient, long-horizon recursive AI workflows.

A protocol and engine for a permissionless compute network that pools VRAM and compute from distributed GPUs to run massive AI models over the open internet.

A low-code REST API platform that automatically creates secure endpoints for databases, enabling faster development of AI and data-driven applications.

A comprehensive training framework for fine-tuning image, video, and audio generative models, featuring a user-friendly web UI and enterprise-grade orchestration.

An LLM inference framework for CPU servers that optimizes memory usage to outperform GPUs in long-horizon, multi-turn inference and ultra-long context tasks.

A native macOS app for Apple Silicon that provides a user-friendly interface to run LLMs, vision models, and image generation locally using the MLX framework.

OpenVINO Model Server (OVMS) is a C++ inference server that serves LLMs, vision‑language models, and classic deep‑learning models via OpenAI‑compatible REST or KServe gRPC APIs. Optimized for Intel CPUs, GPUs and NPUs, it provides features like continuous batching, streaming, model versioning, hot‑reload, and Prometheus metrics. Deployable via Docker, bare‑metal binaries, or Kubernetes, it supports many model formats (TensorFlow, ONNX, OpenVINO IR, GGUF, etc.) and integrates with standard client libraries.

A comprehensive collection of pre-converted Core ML models and sample apps for implementing on-device AI on iOS and macOS.

A structured visual roadmap and resource collection that helps developers transition to AI application engineering by linking papers, code, and diagrams for deep learning models.

A library of over 30 state-of-the-art neural forecasting models designed for high performance, usability, and efficiency in time-series forecasting.