Open source news — new project releases, community governance, licensing updates, and open-source AI models including Llama, Mistral, and others.
A developer building an x86_64 kernel named TaterTOS64 found that AI documentation tools hallucinated incorrect code dependencies and failed to understand the project's repository structure. In response, they built Tater…
The author successfully fine-tuned a 1.1 billion parameter TinyLlama model using QLoRA on consumer hardware, training only 0.2% of the model's parameters via low-rank adapter matrices. The project involved overcoming cha…
Creation of "Local Gemma 4 Content Radar," a privacy-focused editorial tool that runs the Gemma 4 AI model locally to analyze content signals and produce structured publishing decisions. The tool processes a batch of can…
InfraPulse, an open-source tool that simplifies deploying a production observability stack—including Prometheus, Grafana, Loki, and Alertmanager—by embedding all configuration into a single Go binary, eliminating the nee…
A developer reduced AI test automation costs by 200-300x, from $0.011 per step to $0.00004, by replacing vision models with a pure-text approach. Instead of sending full-page screenshots to multimodal LLMs, the new frame…
Creation of Coda AI, a VS Code extension designed to eliminate the workflow friction of constantly switching between a code editor and a browser to use AI tools. Unlike basic AI assistants, Coda understands the entire pr…
The author successfully ran MNIST digit recognition on an ESP32-C3 microcontroller by compiling a neural network directly into a C header file and executing it as firmware, without using TensorFlow, TFLite, or any ML run…
Anthropic's prompt caching often fails silently due to misplaced breakpoints, prefix drift between calls, short TTL expiration (reduced from 1 hour to 5 minutes), and lack of measurement. The author, a solo developer, cr…
"second-brain-cloudflare," a self-hosted MCP server that provides persistent memory for AI assistants like Claude and ChatGPT across sessions, running entirely on Cloudflare's free tier. It uses vector embeddings and a t…
MoonieCode, a minimal AI coding assistant built in 393 lines of C++23 that communicates with Claude via HTTP requests. Its core logic is a single while loop that sends user prompts as JSON to an LLM API, then handles two…
The open-source tool Promptfoo, acquired by OpenAI in March 2026, maps its 155 attack plugins to the OWASP LLM Top 10 2025 list for structured red teaming of LLM-powered products. It details the 2025 revision's new categ…
"Your Story," a Flutter mobile app that uses local Gemma 4 AI models to power interactive fiction. Unlike traditional choose-your-own-adventure games, the app allows players to type natural language actions, which the AI…
An open-source, browser-based AI background removal tool that processes images entirely on the client side, ensuring user privacy. The system uses the `@imgly/background-removal` library with an ONNX model and WebAssembl…
A2UI v0.9 is a new, framework-agnostic standard for generative UI that allows AI agents to dynamically create tailored interface widgets in real-time using a common language, working across web, mobile, and other platfor…
Genkit, an open-source framework for building AI-powered applications, has introduced a middleware system that allows developers to intercept and customize generation calls and tool execution loops. The middleware, avail…
KTransformers is a GitHub project with over 17,000 stars that enables running massive 671-billion parameter models like DeepSeek-R1 on a single consumer machine (e.g., with 512GB RAM and one RTX 4090) at 286 tokens per s…
The Google AI Edge Gallery app now supports the open-source Model Context Protocol (MCP), allowing on-device Gemma 4 models to dynamically call external tools and data sources via Streamable HTTP. The update also introdu…
KTransformers is an open-source inference library that enables running large language models with up to 671 billion parameters on a single consumer-grade machine by using heterogeneous placement to split computation betw…
As engineering teams iterate on prompts for large language models, they face challenges similar to managing configuration files, requiring dedicated storage, versioning, sharing, and security. It reviews existing solutio…
Platform engineering should leverage existing tools like GitHub rather than building separate portals, as developers already spend most of their time in GitHub. It outlines a strategy using "golden-path" starter repos wi…