Tag
An introduction to bb, an open-source, MIT-licensed agentic IDE that functions as an agent orchestrator and can be extended by users to build custom features, including task management, code review, and more.
CL-Forth is a Common Lisp implementation of the Forth 2012 Standard, supporting multiple Lisp implementations and operating systems, with a standalone executable option.
Ant Group's inclusionAI released Ling-3.0-flash, a 124B-parameter (5.1B active) model under a plain MIT license with no revenue caps or usage restrictions, and it's available on OpenRouter with a free tier.
AntLingAGI announces Ling 3.0 flash, a native hybrid-linear reasoning model with 124B total parameters and 5.1B active, MIT-licensed, claiming to match a 1T flagship with much less compute and faster response times.
WebRover is an MIT-licensed open-source AI agent that uses natural language to drive the browser for web automation, cross-site data scraping, and deep research, with support for local deployment.
Meituan released LongCat-Flash-Lite-Sparse, a sparse-attention model supporting 1M-token context with only ~3B active parameters per token, achieving strong SWE-Bench scores under an MIT license.
BrowserSmith 1.0 is released, an open-source tool that orchestrates four ChatGPT tabs to plan, write, review, and audit a project, then execute it, all without requiring an API key.
hwatu is a lightweight verification browser for local coding agents, featuring headless WebKit, pixel-diff scoring with real match percentage, and no Chromium dependency; it is MIT-licensed and written in Rust.
Domyn-Small is a 10-billion-parameter open-weight reasoning language model released under MIT license, trained on 9 trillion tokens and optimized for reasoning, instruction following, and tool use. It achieves strong accuracy-efficiency balance against peers like Qwen3.5-9B and OLMo-3-7B-Think.
An open-source alternative to CapCut called OpenCut has been released, offering a free video editor with no watermarks, paywalls, or subscriptions, featuring a plugin-based architecture and MIT license.
Autonomous open-sourced complete build guides for a personal AI computer with up to 384GB VRAM, enabling local running of open models with no API costs or service shutdown risks.
Fenris Creations has open-sourced the Carbon engine used in Eve Online, making it freely available on GitHub under permissive licenses to build community trust and enable contributions.
Frugon is a free, local, open-source LLM cost analyzer that helps developers find which LLM calls could be handled by cheaper models, reducing costs by analyzing logs and providing routing recommendations.
Kyutai released Pocket TTS, a text-to-speech model capable of cloning a voice from just 5 seconds of audio, running on CPU and released under the MIT license. It was benchmarked against Kokoro, Supertonic, and Inflect-Nano for English TTS.
LongCat-2.0 from Meituan is now fully open-sourced under MIT license, with model weights and inference code released.
Someone reverse-engineered and open-sourced Claude Design's system prompt, providing a design philosophy and prompt library with a 20-chapter main prompt and 14 skills to improve LLM output quality in UI design.
TraceSage is a new MIT-licensed, local-first observability tool for LangChain and LangGraph that runs entirely on your machine, providing live topology graphs, step-by-step replay, token tracking, and OpenTelemetry export without sending data to the cloud.
An open-source, self-hosted AI gateway providing a single endpoint for 237 LLM providers with auto-fallback, token compression, and routing. It has gained significant traction with 9.8K GitHub stars and 280+ contributors in 4.5 months.
ActionProof is an MIT-licensed library that generates tamper-evident cryptographic receipts for AI agent actions, enabling offline verification without backend infrastructure. It supports TypeScript, Python, and drops into Claude Desktop Cursor as an MCP server.
Ornith-1.0 is a family of open-source, self-improving models for agentic coding, achieving state-of-the-art performance on coding benchmarks via reinforcement learning that jointly optimizes scaffold and solution rollouts.