Local & Self-Hosted AI Tools

YouDub-webui is a self-hosted video translation and dubbing app for creators and small teams making videos available in another language. Its most established workflow is English-to-Chinese dubbing for YouTube; it also supports Chinese-to-English dubbing for Bilibili.

Whispering Tiger creates live transcripts and translations from audio on your computer for VRChat conversations and streaming captions. It can also read text in game images, so spoken dialogue and visible text can feed into the same translation workflow.

Taskyon is a browser-based AI workspace for people investigating technical problems and developing engineering designs. It keeps the work in a task tree so related questions, evidence and decisions stay connected.

Bolna builds voice AI agents for businesses handling support, sales and recruitment calls. It focuses on Indian languages and accents, including Hindi, Hinglish, Tamil and Telugu, alongside English.

KanDev is a self-hosted workspace for developers managing AI coding tasks and reviewing the resulting changes. It organizes work on kanban boards and lets you define workflows with your own agent choices and review gates.

Vexa is a meeting bot and transcription API for developers adding meeting features to products and teams connecting calls to AI agents. Its bots join Google Meet, Microsoft Teams and Zoom to capture the conversation.

Lexora is a personal desktop AI agent for research, creative projects and everyday planning as well as coding. It works toward a goal while you give direction and review its work.

GraphRAG is an open-source Python indexing and retrieval pipeline for developers building question answering over their own document collections. It can connect to local models through Ollama proxies, while OpenAI models are its default supported choice. With a remote model provider, text sent for model processing leaves your machine.

DeskcommCRM is a self-hosted CRM for businesses selling through WhatsApp. Its AI agents handle customer conversations, qualify leads and support sales within the CRM.

Parakeety is a paid Mac dictation app for people who want to speak into the app they're already using without sending their voice to a cloud service. It runs on Apple Silicon Macs with macOS 14 or later and places the transcript at the cursor. Intel Macs aren't supported.

bitsandbytes is a Python library for developers who need to run or fine-tune large language models on their own hardware with less memory. It works within PyTorch workflows.

ChatWise is a desktop AI chat app for people who want to use different model providers in one interface. It keeps chat data on your device while sending requests to the provider you choose.

Amical is an AI dictation app for people who want to write emails, messages and coding prompts by speaking. It adjusts the resulting text to the app you're using, reducing the editing needed after transcription.

CodeGen is a family of language models for generating programs, aimed at researchers and developers working on code generation. Salesforce AI Research studies how these models build programs through multiple rounds of prompting.

Kolors generates photorealistic images from text for creators and developers who want to run image generation on their own hardware. Developed by Kuaishou, it supports Chinese and English prompts, including descriptions of Chinese cultural content.

DeepSeek-Coder is a family of coding models for developers who want to run a model on their own hardware or build it into a coding assistant. It includes base models for completion and instruction-tuned models for coding requests written in plain language.

AIRecon automates penetration testing and bug bounty reconnaissance for security testers. It guides an assessment through reconnaissance, analysis, exploitation and reporting, with a terminal interface for following the agent's work.

WhisperJAV generates subtitles for Japanese adult videos on your own computer. It targets difficult audio with quiet dialogue, loud vocalisations and long stretches that can cause speech recognition models to invent or repeat words.

RKLLM is a toolkit for developers building local AI applications on Rockchip boards. It runs language and multimodal models on the chip's neural processing unit (NPU).

faster-whisper is an open source Python library for developers adding speech transcription to their software. It uses CTranslate2 to run OpenAI's Whisper models, with faster transcription and lower memory use than the original implementation in the project's benchmark comparisons.

Step-Audio 2 is an audio AI model family for developers building voice assistants and applications that interpret recordings. It can reason about what people say and how they say it, as well as sounds beyond speech.

Keras is an open-source Python deep learning framework for developers and researchers who want to build and train models without tying their code to one backend. It runs locally on Linux and macOS, with WSL2 recommended for Windows. Its Apache 2.0 license permits commercial use and modification.

Netron visualizes neural network and machine learning models for developers and researchers who want to examine their structure.

Cloud CLI is a browser interface for developers who want to work with coding agents and manage projects from a computer or phone. It connects to Claude Code, OpenCode, Cursor CLI and Codex, so you can access agent sessions beyond the terminal.

qabot lets you ask questions about files and databases in plain language from your terminal. It pairs a local DuckDB database with an AI model that writes and executes SQL, making it useful for analysts and developers who want to explore data without writing every query themselves.

USearch is a vector similarity search library for developers building semantic search and recommendation systems within their own apps or servers. Its compact C++ core supports custom similarity rules and has no mandatory dependencies.

DB-GPT is a self-hosted AI data assistant for teams analyzing business data and developers building data applications. It connects plain-language questions to database queries and code-based analysis.

DSH AgentTeams coordinates AI agents inside DeepSeek Harness for coding, research and review tasks. A lead agent assigns work to specialists and combines their results, while you can review the proposed team before execution.

Paperless-GPT adds AI document processing to an existing paperless-ngx archive. It's for people who want to reduce manual sorting and extract text from scanned paperwork.

Model2Vec is a Python package for developers who need text embeddings for search or classification on their own hardware. It turns sentence transformers into smaller static embedding models, trading a small drop in performance for lower model size and faster CPU inference.

OpenWhispr is a dictation and meeting transcription app for people who want to write by voice and keep searchable records of conversations. It lets you choose local processing or cloud providers.

rerankers is a Python library for developers building search and retrieval systems. It ranks candidate documents against a query and lets you compare reranking approaches through a shared interface.

AutoRAG Agent searches documents and connected archives where they already live, then reads the relevant sources to produce answers with file, page or line references. It runs on macOS, Windows and Linux and suits people researching information spread across local files, messages and notes.

SeedVR is an AI video restoration project for developers and researchers processing real-world footage or AI-generated video on their own GPUs. Its SeedVR2 model restores footage in a single diffusion step, with substantial hardware demands in the supplied implementation.

Ollama Python is a client library for developers adding model conversations to Python applications. It connects to a running Ollama service for local use or to Ollama's cloud service for larger models.

LightEval is a Python toolkit from Hugging Face for developers and researchers comparing LLMs or investigating where their responses fail. It evaluates models already loaded in memory as well as models accessed through a service.

PYTHIA monitors world events and generates forecasts for people tracking global risks and developers building AI agents that need current-world context. It combines Osiris's globe and live feeds with MiroFish's forecasting approach and local Ollama models.

iZotope Nectar is a proprietary collection of vocal processing plugins for producers, audio engineers, musicians, and creators working with singing or dialogue. Its Vocal Assistant recommends a starting point for a mix that you can adjust to suit the recording. The assistant sets EQ, reduces sibilance, and creates space for the voice within the mix.

Oterm is a terminal chat client for people who want to use local LLMs or hosted models without a browser interface. It keeps separate conversations and their settings so you can return to ongoing work.

Grounding DINO locates objects in images from category names or descriptive text. It's for developers and computer vision researchers building detection or annotation workflows that need labels beyond a fixed category list.

Airux Pocket AI runs AI conversations on Android phones for people who want offline chat and photo analysis without cloud inference. You don't need an account, and downloaded models run directly on your phone.

ExLlamaV2 and V3 are local LLM inference libraries for developers building AI applications or hosting models on consumer GPUs. V2 is archived for now and isn't currently maintained; development continues in V3.

Vibe is a desktop app for people who need transcripts or subtitles from recordings without uploading them to a transcription service. It handles audio and video locally, including interviews, meetings and material in different languages.

Data-Juicer is a Python framework for researchers and teams preparing datasets for AI training, evaluation, agents or document retrieval. It processes data on your own machine or across a Ray cluster.

FLUX.1 is a family of image models for developers and creators who want to generate and edit images on their own hardware or servers.

CLIP is an image and text model for developers and researchers who want to classify pictures without building a labeled dataset for each task. You define the categories in natural language, and the model scores how closely each description matches an image.