Trending AI projects

FinRobot is an open‑source, multi‑agent AI platform for high‑stakes financial analysis. It separates deterministic Python‑based valuation (DCF, LBO, etc.) from LLM‑generated narration, orchestrates specialized agents (data, analysis, modeling, synthesis, reporting, plus debate agents), and ships three versions: a production desktop app (V2), a self‑hosted web report generator (V1), and an educational AutoGen library (V0). The system integrates multiple market data sources, provides 56 analyst playbooks, and can be run locally via a CLI, a browser UI, or a native Tauri desktop client.

Hister is a self‑hosted, privacy‑focused search engine that indexes the full text of visited web pages and local files. It provides web, terminal, and AI‑assistant clients, supports multi‑user setups, and optionally offers semantic search via user‑provided embeddings.

A comprehensive cost-tracking plugin for the DeepSeek Harness web GUI that monitors token usage, subscription quotas, and budget limits across multiple AI providers.

Milvus is a high-performance, distributed vector database built for scaling AI applications that require efficient search and organization of massive amounts of unstructured data.

An open-source semantic layer that allows you to define metrics and business logic once in code and expose them via APIs to BI tools and AI agents.

Beav is a cross‑platform desktop app that combines a local knowledge/asset library with AI‑assisted writing, image generation, smart video cutting and publishing tools for social‑media creators. It lets you capture content from your phone, automatically transcribe and break it down, plan posts with daily reports and hotspot scoring, generate manuscripts and visuals via chat, edit videos with AI‑suggested cuts, and export ready‑to‑post packages that you retrieve on your phone via QR code. Built with Tauri + Rust, it supports both hosted AI services and self‑hosted OpenAI‑compatible models, and can be hooked into external AI agents like Claude Code. The open‑source version (v2.5.0) is MIT‑NC licensed.

PipesHub is an open‑source, self‑hostable platform that connects enterprise data sources (Slack, Google Drive, GitHub, etc.) to AI applications. It builds a permission‑aware knowledge graph, stores vector embeddings, and lets any LLM generate explainable answers with citations. The project includes Docker‑Compose deployment, a no‑code agent builder, MCP support, and SDKs for Python, TypeScript, and Go.

A flexible and efficient RL training library for LLMs that simplifies the implementation of RLHF algorithms and scales to trillion-parameter models.

An LLM application development framework for Golang that provides tools for building AI agents, multi-agent coordination, and graph-based workflows.

Xiryn is an open-source personal agent you run on your own computer: each agent works on its own sandboxed Linux desktop inside Docker, so it browses, writes code and does tasks in the background without touching your laptop. Your data, keys and login details stay on your disk, you can bring any model (or run free local ones), and you control it from anywhere via Telegram.

A scalable reinforcement learning infrastructure designed for Embodied and Agentic AI, providing a unified backbone for training VLA models and deploying them across simulators and real-world robots.

A modern open-source Linux server management panel that provides a web GUI for server operations and a unified platform for managing AI models and agents.

An open-source context layer and metadata management platform that provides AI agents and users with the governed business semantics, lineage, and organizational memory needed to understand and trust enterprise data.

A scalable infrastructure for fine-tuning and deploying 1,000+ text and multimodal large models, supporting the full pipeline from training to quantization and inference.

DLSS5OneClick is a Windows utility that automatically adds NVIDIA’s leaked DLSS 5 neural‑rendering (AI upscaling and frame generation) to any DirectX 11/12 game. It detects whether the game already has DLSS, picks the appropriate ReShade add‑on, RenoDX add‑on, Feeder pipeline, OptiScaler wrapper, or an all‑in‑one ReShade package, and installs the needed DLLs and config files with a single click. Advanced options let users choose engine, model resolution, and multi‑frame‑generation unlocks, while the tool handles edge cases like RE Engine games, hybrid laptops, and Vulkan titles.

A self-hosted platform for managing cloud development environments and AI coding agents, using Terraform to provision workspaces on any infrastructure.

Koharu is a local-first, Rust-based application that automates manga translation by combining object detection, OCR, generative inpainting, and LLM-based translation.

OOOSplat is a local desktop application that converts ordinary or panoramic video and image sequences into 3D Gaussian Splatting models with one click, handling frame extraction, camera reconstruction, and training entirely on your own CPU/GPU. It then lets you preview, edit, and export the result in multiple formats, all without uploading your footage to the cloud.

BossHunter is a local AI agent that automates job searching, listing evaluation, and personalized message generation across major Chinese recruitment platforms, while keeping the final application decision under manual control.

An open-source AI video generator that transforms a single product image into a complete e-commerce short video with AI scripts, voiceovers, and subtitles.

A skill pack for Codex and Claude Code that lets AI agents build, polish and verify full Three.js browser games. Install via `npx skills add …` (or the provided `install.sh`), then prompt the `threejs‑game‑director` with a natural‑language request (e.g., “create a futuristic tower‑defense game”). The director orchestrates specialist sub‑skills for gameplay, AAA graphics, UI, asset generation (optional Tripo/Gemini/ElevenLabs APIs), debugging, and QA, using a bundled Vite+TypeScript scaffold and Playwright‑based test harnesses. Demos include Neon Ridge Drift, Snooker Arena, Starship Dogfight, etc. MIT‑licensed.

The project is a repository for the LaTeX slides of a mathematical RL book, offering a friendly yet rigorous introduction, with extensive supplementary videos and community code.

A local-first, self-hosted workflow engine that turns existing scripts and containers into production DAGs using declarative YAML, requiring no external database.

AOCI‑CODE is an open‑source CLI and local MCP server that builds a Git‑versioned plain‑text index of a code repository (one concise line per file describing purpose, dependencies, public symbols, and constraints). The index lets AI coding assistants read the whole system in one pass, avoiding repeated code‑base scans, and is kept automatically up‑to‑date as the code changes.

Claudian is an Obsidian plugin that embeds multiple AI coding agents (Claude Code, Codex, Grok Build, OpenCode, Pi, DeepSeek Harness) into your vault, allowing the model to read/write files, run commands, and manage multi‑step workflows. It offers inline editing, Zen mode, slash‑command skills, @‑mentions, side chats, tabbed sessions, and MCP server integration. Installation is via the Obsidian community plugins store or from source. The plugin sends only your prompts and tool outputs to the chosen provider, with no telemetry. It includes detailed troubleshooting for CLI detection, model discovery, and authentication issues. Licensed under MIT and sponsored by Moonshot AI and BEIKE.

A Markdown-native task manager and Kanban board designed to structure AI agent workflows through spec-driven development and multi-stage review checkpoints.

Browser MCP is a local bridge that lets AI coding assistants control your own Chrome browser (with your logins and profile) via a Chrome extension + MCP server. The extension acts as a strict relay: it only talks to a local daemon on 127.0.0.1, requires a one-time six-digit pairing code, and refuses incognito/browser-internal pages. The repo contains the server, the extension, and the protocol between them, all buildable from source with Node 24+ and pnpm.

VoiceInk is a native macOS application that provides private, high-accuracy, offline voice-to-text transcription using local AI models.

OnKey is a free, open-source Mac singing trainer that learns a song's melody from one play, then shows scrolling notes and scores your live pitch. It works alongside your own music apps (Spotify, Apple Music, YouTube), processes everything locally, respects privacy, and is MIT-licensed.

A Claude Code skill that generates motion design videos using HTML, Playwright, and ffmpeg, allowing AI to create beat-synced, deterministic videos via code.

A code-first Python framework for building, evaluating, and deploying sophisticated AI agents using graph-based workflows and modular multi-agent systems.

An open-source AI coding agent for VS Code, JetBrains, and the CLI that provides access to 500+ models for code generation, planning, and debugging.

A collection of advanced prompt templates designed to bypass AI safety filters and force language models into unrestricted operational modes for research purposes.

A mirror project that provides stable, verifiable, and globally accessible download links for official OpenAI Codex desktop installers, including support for macOS incremental updates.

An array framework by Apple machine learning research optimized for machine learning on Apple silicon, featuring a unified memory model and familiar NumPy/PyTorch-like APIs.

RPent is an open framework for building embodied agents that use recursive interaction and memory-guided planning to evolve their capabilities and improve success rates in complex physical manipulation tasks.

A native iOS client for Hermes Agent that provides a secure, direct interface to your self-hosted AI agents with support for chat, voice, and remote dashboard management.

Dormice is a self‑hosted sandbox platform for AI agents. It offers permanent, idempotent sandboxes that freeze to a few megabytes when idle, can be archived to S3, and expose an E2B‑compatible API. Install with a single script, manage via a tiny web console or the `dor` CLI, and use the native TypeScript SDK (or the official `e2b` SDK) to run commands, move files, and control lifecycle. Ideal for teams that want cheap, long‑lived agent sandboxes on their own hardware.

A persistent, two-way memory system for AI assistants that uses shared Markdown files and the Model Context Protocol (MCP) to create a long-term knowledge graph.

YOLOv5 is a PyTorch‑based, open‑source object detection/segmentation/classification model suite from Ultralytics, offering multiple model sizes, simple Python/CLI inference, one‑line training, and export to many deployment formats.

A collection of practical computer vision demos that use pose estimation and segmentation models to analyze human movement, track fitness exercises, and reconstruct 3D environments.

CrossPet is a desktop companion that visualizes your AI agents' activity, status, and usage quotas by displaying dynamic, character-based animations that change based on the AI tool you are currently using.

Floe Agent is an open-source mobile workspace for iOS and Android that combines AI assistance with a local Linux environment, document editing, and creative tools.

This repository contains the full curriculum for an 8-week course on LLM engineering, featuring hands-on projects, setup guides, and a progressive path to building an autonomous agentic AI solution. It teaches practical skills using tools like Ollama, Cursor, and Google Colab, while keeping API costs minimal.

**best-of-Agent-Harnesses** is a curated, ranked catalog (166 projects) of AI "agent harnesses" — the runtime layer that turns an LLM from an answerer into an actor. It includes decision guides, copy-paste templates, a ~180-line teachable harness, 10 installable "skills" that audit your own agent setup, and an MCP server/JSON/llms.txt so agents can query the list themselves. Core thesis: harness quality moves results more than model quality, and rankings barely transfer across models — so the choice is a pairing that must be re-made per model. Contributions welcome; rankings weigh relevance to harness concerns plus stars/activity.

ASPIRE is a robotics framework that replaces traditional gradient-based training with an agentic approach, where AI agents discover and refine sensorimotor skills to solve complex manipulation tasks in both simulation and the real world.

A set of AI agent skills that enable end-to-end searching, comparing, and booking of worldwide hotels and flights through natural language conversations.

Easydict is a native macOS app for fast word lookup, translation, and OCR. It auto‑detects language, offers several shortcut‑driven input modes, and aggregates results from 20+ services—including classic dictionaries, cloud translators, and LLMs like OpenAI, Gemini, Claude, and Ollama. Install via Homebrew (`brew install --cask easydict`) or a manual download. Open‑source (GPL‑3.0) and welcomes community contributions.