AI & LLM Tooling

Open-source agentic assistant, UI-first, no config-first setup.

Lightweight TTS app from Kyutai Labs designed to run efficiently on CPUs — 100M params, ~236MB, ~200ms latency, 6x realtime on CPU, multi-language, voice cloning, can run client-side in browser.

Open-source, self-hostable voice agent — separates the AI 'brain' from the physical device so any board with mic/speaker/WiFi can act as the body. Fully documented hardware/software connection.

Claude Code skill that runs an After Action Review after any task/session — pulls concrete lessons into a persistent LESSONS.md file so the agent stops repeating the same mistakes across sessions.

AI skill providing design intelligence for building UI/UX.

Open-source AI research agent CLI — audits papers against public codebases, runs autonomous research loops, produces cited literature reviews and polished paper-style drafts. Built on the Pi agent runtime.

Pure-C, zero-dependency inference engine that runs frontier MoE models (GLM-5.2 744B, Kimi K3 2.8T, etc.) on consumer hardware with ~25GB RAM by streaming experts from disk/SSD.

Unified OpenAI-compatible endpoint routing across 34 free-tier LLM providers (Gemini, Groq, NVIDIA NIM, OpenRouter, etc.) with smart routing and automatic failover; ~7.4B tokens/month pooled.

AI tool to create interactive presentations from a prompt.

Rust terminal tool that scans your hardware (CPU/RAM/GPU) and scores hundreds of local LLM models for fit, speed, quality and context so you know what will actually run before downloading.