Local AI

VoiceStudio

debpalash/VoiceStudio

Huia Project Score 80

VoiceStudio is a desktop app that runs speech AI on your own computer: it clones and designs voices, generates speech from text, dubs videos, transcribes audio and has a dictation widget. The app is built on Electron with a Python backend. Its default speech engine is k2-fsa's OmniVoice, and the feature catalog lists more than a dozen alternatives, among them CosyVoice 3, KittenTTS, MLX-Audio and IndexTTS 2.5. Transcription defaults to WhisperX, with Faster-Whisper, Parakeet, Moonshine and others available. Models are downloaded when a workspace first needs them. The project says its tools cover 646 languages, a figure that also appears in the transcription tool of its MCP docs. Beyond the desktop workspaces, VoiceStudio exposes a local API and an MCP server on port 3900, so coding agents can generate speech, clone a voice or transcribe a file. The README states that local workflows stay on your hardware, remote workers are optional and usage analytics only run with consent. Version 0.5.3 was the last release built on Tauri; current releases are Electron only, and Tauri users have to install the new app separately.

Stars

61K

Language

Python

License

AGPL-3.0

Momentum

+16.0% 7dTrending

omlx

jundot/omlx

Huia Project Score 73

oMLX is a local LLM inference server for Macs with Apple Silicon, built on MLX and managed from a menu bar app. It exposes OpenAI-compatible and Anthropic-compatible endpoints, so existing clients can connect to http://localhost:8000/v1 without code changes. Its main idea is a two-tier KV cache. Frequently used cache blocks stay in RAM, and when that hot tier fills up, blocks are offloaded to the SSD in safetensors format so they can be reused instead of recomputed. The author pitches this at long coding sessions with tools such as Claude Code. Concurrent requests go through continuous batching using mlx-lm's BatchGenerator. The server loads text LLMs, vision-language models, OCR models, embeddings and rerankers. Several models can stay loaded at once, with least-recently-used eviction, pinning and a per-model idle timeout. A web admin dashboard at /admin covers monitoring, model management, chat and benchmarks. Source builds also include an experimental mode that splits one model across several Macs. You can install it as a .dmg from GitHub Releases, through Homebrew, or from source, and it is licensed under Apache-2.0.

Stars

23K

Language

Python

License

Apache-2.0

Momentum

Established

llmfit

AlexsJones/llmfit

Huia Project Score 72

llmfit is a terminal tool, written in Rust, that inspects your hardware and tells you which open-source large language models it can run well. It detects CPU cores, system RAM, GPUs, VRAM and unified memory (NVIDIA CUDA, Apple Silicon, AMD ROCm and Intel OneAPI), then estimates memory use and tokens per second for each model across quantizations. Each model gets scores for quality, speed, fit and context, plus a fit verdict that runs from Perfect to Too Tight. The default interface is an interactive TUI. A classic CLI mode, a web dashboard and a REST API (/api/v1/system and /api/v1/models) are also available. llmfit works with Ollama, llama.cpp, MLX, Docker Model Runner and LM Studio, and can download a chosen model through them. Speed figures come from a memory-bandwidth formula unless a measurement exists. A newer benchmark feature measures real tok/s on your machine and lets you submit the result to the project as a pull request from the TUI, so later releases carry measured numbers for matching hardware. The tool runs on macOS, Linux and Windows and is MIT licensed. The latest release is v1.1.17.

Stars

38K

Language

Rust

License

MIT

Momentum

Established

DigUp

ARahim3/DigUp

Huia Project Score 70

DigUp is a free macOS menu bar app that searches the files on your Mac by what is in them, using Google DeepMind's EmbeddingGemma 2 model running locally. You pick the folders it may read. DigUp then indexes pictures, screenshots, PDFs, documents, audio recordings and videos, and you find things by typing a plain description. A hit in a video or recording opens at the matching moment, and a PDF hit opens at the page. Exact strings such as invoice numbers and error codes are matched too, including text that Apple's Vision framework reads from screenshots. Code search is a separate index you switch on yourself: queries start with code: and results open at the line in editors such as Zed, VS Code, Cursor and Xcode. The README says nothing you index leaves the Mac. The app goes online to download the model once (865 MB) and for a daily update check that can be turned off, and it has no account and no telemetry. Internally it runs EmbeddingGemma 2 on llama.cpp with Metal and keeps vectors plus a keyword index in one SQLite file. It needs an Apple Silicon Mac on macOS 14 or later. The latest release is DigUp 0.6.0.

Stars

1.5K

Language

Swift

License

MIT

Momentum

Emerging

mempalace

MemPalace/mempalace

Huia Project Score 69

MemPalace is a Python memory layer for AI assistants: it saves your past conversations and project files word for word on your own machine, then lets you or your agent search them by meaning. The project deliberately avoids summarizing what it stores. Records are filed into a hierarchy the authors call a palace: each person or project gets a wing, subjects inside it become rooms, and the original text sits in drawers. A search can then stay inside one wing instead of scanning everything. ChromaDB is the default vector store, and the repository ships opt-in adapters for SQLite, a native Rust engine, Milvus, Qdrant and pgvector. Day to day, you point the mempalace CLI at a code folder or at Claude Code session logs, and it indexes them. An MCP server exposes the palace to agents with 45 tools, and hooks for Claude Code, Codex CLI and Cursor save sessions on a schedule and before context compression. The README reports 96.6% recall at five on the LongMemEval benchmark using plain semantic search with no language model involved, and says the scripts to reproduce that figure are in the repository.

Stars

60K

Language

Python

License

MIT

Momentum

+0.2% 7dEstablished

nanobot

HKUDS/nanobot

Huia Project Score 69

nanobot is a self-hosted personal AI agent framework in Python from HKUDS. You run it on your own computer or server and talk to it through a browser WebUI, the terminal, or chat apps such as Telegram, Discord, Slack, WeChat, email and Mattermost. The agent can read and write files, run shell commands, search and fetch web pages, call MCP servers, schedule jobs with cron, generate images and hand work to subagents. Session history and long-term memory are handled by a component the project calls Dream. Models come from OpenAI-compatible APIs or local servers such as Ollama and vLLM, with fallback models configurable. nanobot can also expose its own OpenAI-compatible API, and it has a Python SDK for integrations. The stable track is the nanobot-ai package on PyPI, installed with a one-line script, uv or pip; the latest release is v0.3.5. Newer and experimental features land on the main branch first and need a source install. Documentation is at nanobot.wiki in ten languages.

Stars

49K

Language

Python

License

MIT

Momentum

+0.5% 7dEstablished

turbovec

RyanCodrai/turbovec

Huia Project Score 65

turbovec is an in-process vector index for Python and Rust that stores embeddings at 2 or 4 bits per dimension and searches the compressed vectors directly. It is written in Rust with Python bindings and uses TurboQuant, a quantization method from Google Research. The README's headline figures: ten million 768-dimensional embeddings that need 31 GB as float32 fit in 4 GB, and search runs 4.4× faster than FAISS FastScan at 4-bit, at matching recall. Its benchmark table puts the 4-bit speedup at 3.8–5.1× across eight test cells on GCP ARM and x86 machines. These are the project's own benchmarks, and the scripts and raw JSON are published alongside them. No training step is needed, so a vector is searchable as soon as it is added, and removal takes microseconds. sync() writes only what changed since the last call. IdMapIndex keys vectors by your own uint64 ids and accepts an allowlist filter at query time. Drop-in stores exist for LangChain, LlamaIndex, Haystack and Agno. It is a library rather than a server, and it scans every vector instead of using a graph index.

Stars

17K

Language

Rust

License

MIT

Momentum

Established

An open-source tool and API for running open models on your own machine or on Ollama's cloud.

Desktop apps for running open LLMs locally, with a local API server and Bionic, an agent for documents, code and automation.

Push-to-talk dictation for the Mac that types where your cursor is, plus an AI helper that points to controls on screen.

Mac app that keeps an encrypted, searchable memory of your screen and meetings, with local MCP access for AI coding agents.

A TypeScript framework for AI apps that run an open-weight model inside the app, offline and without an API key.

An 11 MB open-weight model from HITL Systems that answers routing, yes/no and priority questions about text on CPU.

Mac app that runs open AI models up to 35B parameters offline for chat, transcription, OCR, upscaling and video clipping.

An iOS and Android app that runs 53 open-weight models offline on your phone, with no account or subscription.