local-ai

A free, living knowledge base on vibe coding — building SaaS fast with AI agents, local AI, and agentic workflows.

.NET library for local embedding generation using ONNX and Microsoft.Extensions.AI

Eve Agent V2 Unleashed — local-first autonomous AI coding agent powered by Ollama

The highest-scoring AI memory system ever benchmarked that isn't reliant on LLM reranking. And it's free & burns less tokens.

Local, private push-to-talk dictation for macOS. Hold a key, speak, release — your words appear in any text field. Runs entirely on-device via whisper.cpp with bundled multilingual model. No cloud, no telemetry, no paywall.

Universal local LLM runtime. Drop-in Ollama replacement with HuggingFace integration, pipelines, structured output. Free & open source.

MeetingBro — Open-source, local-first AI meeting assistant. Real-time transcription, translation & live summaries. No platform plugin needed. Windows · macOS · Linux

Powerful no-code LLM fine-tuner: upload data → train → deploy in minutes. Unsloth 2-5× acceleration · QLoRA/DPO/RLHF/PPO/ORPO · Reward Model training · GGUF export · vLLM inference · BLEU/ROUGE/BERTScore · full CLI · Heretic Mode to unlock full model potential

😼 OpenFelix — Voice-first AI agent for macOS. Runs local LLMs on Apple Silicon via MLX. Supports all OpenClaw skills. Claude, GPT, or fully offline.

StickerSelector 是一个基于语义向量匹配的表情包选择系统，用于让 AI 在聊天中发送真正符合语境的表情包。 与传统依赖关键词或规则的方案不同，StickerSelector 会将 AI 生成的“表情包意图描述”与本地表情包进行语义匹配，从而选出在当前语境下最自然、最像真人会使用的表情。 适用于 AI 聊天机器人、QQ或微信的自动聊天、AI 女友等需要“拟人化表达”的场景。

Your LLM playground on Apple devices. Run open-source AI models locally on your Mac with a native SwiftUI interface.

An advanced Self-Centered Intelligence (SCI) prototype that represents a new paradigm in AI-human interaction.

Self-hosted, OpenAI-compatible speech server: faster-whisper STT and Piper/Kokoro TTS. A local voice backend for Open WebUI, CLIs, and custom agents.

DRIA (Deep Research and Intelligence Agent) is a fully local voice assistant that can hold real-time conversations while performing deep research in the background — powered by Firecrawl, Mistral AI, Perplexica, and LiveKit.

Blueprint by Mozilla.ai on how to transcribe audio files

Unified KV-cache compression for LLM inference: 12 Python-native methods, guarded add-on composition and routing, analytical capacity simulation, Godzilla KVarN/TriAttention with SM86/SM89 qualification, CUDA weight sharing, and multi-GPU planning.

Local AI workstation — discover, run, chat, benchmark, and generate images from open-weight models. DFlash/DDTree speculative decoding, TurboQuant & TriAttention cache compression strategies, MLX + llama.cpp + vLLM + MTPLX backends.

Semantic search knowledge base for MCP-enabled AI assistants. Index local files or GitHub repos, query with natural language. Built on LanceDB vector storage. Works with Claude Desktop, Cursor, and other MCP clients.

An open-source, model-agnostic agent harness for local LLMs. Define agents in YAML (tools, memory, deny-first permissions) and run them against any OpenAI-compatible endpoint: vLLM, Ollama, LM Studio, or llama.cpp.

High-performance, Hardware-First Vector Database engine built in Zig and Go. Sub-millisecond HNSW search with 64-bit metadata filtering and 1.8MB binary footprint.