Local & Self-Hosted AI Tools

Agent Service Toolkit is a Python template for developers building their own AI agent services. It brings together agent logic, an API and a browser chat interface so you don't have to build each part from scratch.

Remembra Relay saves coding session handoffs so developers can switch assistants or resume a project on another machine without reconstructing its progress. The next agent receives a brief of completed work, unfinished tasks and failing checks.

Whisperstream turns speech into text in the app you're writing in. It's for people who want to dictate messages, documents or coding prompts while keeping speech recognition on their computer.

Scira is an AI search engine for researching questions across live sources and checking the evidence behind an answer. You can self-host the application, but live research needs internet access and external services.

GLaDOS is a local AI voice assistant for people who want a conversational companion modeled on the personality core from Valve's Portal games. It combines a custom voice and dry humor with awareness of its surroundings.

gpustat is a local terminal monitor for people running AI workloads or sharing accelerator hardware. It helps you check device activity and identify the jobs using memory without opening a separate dashboard.

Skyvern automates browser tasks for teams handling forms, vendor portals and repeated data collection. It uses vision models to identify page elements, helping workflows adapt when website layouts change.

Agenta is an AI agent workspace for teams that want to delegate work and manage recurring tasks together. Agents plan their own steps, use connected tools and carry feedback into later work.

Dictus Desktop is a free, open source voice dictation app for macOS, Windows and Linux. It turns speech into text in the app you're typing in, with transcription handled entirely on your computer. It's for people who want to dictate messages, documents or emails without sending their recordings to a cloud service.

Priv AI is an on-device assistant for people who want to chat with AI and use it to examine personal health and financial information. It works offline after a model download and doesn't require an app account.

PersonalClaw is a self-hosted AI agent for one person who wants chat, ongoing tasks and personal knowledge in one web dashboard. It runs on Linux and macOS, with Docker support that also covers Windows. It's open source under MIT and is currently pre-1.0, with breaking changes expected. A one-time download supplies a tiny CPU model for offline chat without tools. Autonomous tool work needs your own Ollama model or a provider app.

Albert coordinates Claude Code agents for developers who want to hand off a coding goal that spans several tasks. It plans the work and continues through successive rounds until it reaches the goal or a requested review checkpoint.

Amphion is an audio generation toolkit for researchers and engineers developing speech and singing systems. It brings model implementations, dataset preparation and evaluation into one research codebase.

RSSBox helps people follow foreign-language news through translated RSS feeds. It combines translation, article filtering and daily digests for readers who want to manage multiple sources in one self-hosted service.

4M is a framework for researchers and developers who want to train and run models that combine images, text and other visual information. A single model can perform vision tasks and generate images from mixed or partial inputs.

AI Documentation Generator analyzes repositories and writes documentation for developers who need to understand an unfamiliar codebase or prepare it for new contributors. It's open source under the MIT license and runs as a Python application, in Docker, or through a Claude Code plugin.

StyleTTS 2 is a text-to-speech model for developers and speech researchers who want to generate expressive speech on their own hardware. It can choose a speaking style from the text without a reference recording.

Hollama is an LLM chat interface for people who want to use Ollama or OpenAI servers through a browser. It supports multiple server connections, so you can keep access to different model servers in one app.

PaddleClas is a PaddlePaddle toolkit for developers and researchers building image classification and recognition applications on their own hardware. It covers fixed-category classifiers and image matching for large or frequently changing catalogs.

MiniCPM-V is a family of local vision-language models for developers building applications that interpret images and video. It can run on phones and other hardware you control.

DiffSinger generates synthetic singing voices for voice developers and music production. OpenVPI maintains this fork of MoonInTheRiver/DiffSinger, with acoustic models and adjustable vocal expression controls.

dstack is a self-hosted AI compute orchestration tool for teams running development, training and inference across cloud providers and their own servers. It gives those workloads a shared interface across Kubernetes, VMs and bare-metal clusters.

Agentgateway is an open-source proxy for platform teams connecting applications and agents to models, tools and other agents. It brings routing and access controls together across AI protocols and ordinary service traffic.

Arkitekt OpenCRM is a free, self-hosted CRM for small businesses that want lead management and email outreach on their own server. It runs in Docker with a browser dashboard and stores lead records in SQLite on the host. AI email writing is optional and can use local models through Ollama or cloud services such as OpenAI and Anthropic.

Ollama JavaScript is a library for developers adding model chat, generation or retrieval features to JavaScript applications. It connects those applications to an Ollama server, including one running locally.

Arize Phoenix helps developers investigate failures in AI agents and LLM applications and compare proposed fixes. It connects recorded application activity with evaluations and repeatable experiments.

CleanRL is a Python deep reinforcement learning library for researchers and developers who want to study or modify an algorithm without tracing it through a large modular codebase. Each algorithm variant lives in one standalone file. Experiments can run locally, while Docker and AWS integration support cloud execution through AWS Batch.

Vela is a desktop AI writing editor for novelists and web fiction authors who need to manage a long story's characters, settings and chapters. It brings planning, drafting and revision into one workspace.

Nextcloud Recognize is an AI media tagging app for people organizing photo, video and music collections on a self-hosted Nextcloud server. It processes images on that server and doesn't send sensitive data to cloud recognition services.

Sentence Transformers is a Python library for developers building search and document retrieval on their own hardware. It helps match content by meaning and rank results against a query.

Cherry Studio is a desktop AI chat app for people who use several model providers and want one place to manage their conversations. It supports local model connections as well as cloud services.

Instant NGP is neural graphics software for researchers and developers working on 3D reconstruction and rendering. It trains representations of scenes and images on a single GPU, then lets you explore the results interactively.

OpenBrowser is a self-hosted browser automation agent for people who need visual judgment and interaction across web pages. Its server and Chrome extension run locally, but the documented Qwen setup sends model requests to DashScope's cloud API and requires an API key and internet access. Browser execution stays local; model processing uses the remote provider.

Jamba is AI21's language model family for teams building document-based assistants and business applications on their own servers. The Large model targets lengthy context, but its hardware requirements put it in server territory: quantized deployment uses eight GPUs with 80GB of memory each.

Marqo Open Source is a self-hosted search engine for developers building document search, product discovery, or retrieval for AI applications. The project is deprecated and no longer receives updates.

CatBoost is an open-source machine learning library for developers and data scientists building prediction models with numerical and categorical data. Its built-in handling of categories reduces the work needed to turn non-numeric inputs into usable features. It runs computations on your own CPU or GPU and uses the Apache 2.0 license.

ccusage is a local command-line tool for developers who want to see how many tokens their coding agents use and what that usage costs. It reads the agents' local usage records, so you can compare activity across tools in one place.

MemFree is an open-source AI search tool that answers questions using both your own knowledge base and the internet. It's for people who want to search saved documents, bookmarks and web content together, with AI summaries instead of reading each result separately. You can self-host it or use the hosted website.

exo is a local LLM runner for people who want to run models too large for a single machine's memory. It joins devices into a cluster and runs inference on your own hardware.

Claw Orchestrator gives developers one interface for multiple coding agents and checks their results independently. You define acceptance criteria, and the runtime executes the checks and keeps the evidence on disk.

EverOS is a Python library and memory runtime for developers building agents that need context across sessions and applications. It keeps conversations, files and task histories in readable Markdown that you can edit and version with Git.

Swarm Fleet manages multiple terminal AI agents on your machine for developers coordinating coding, review and ticket work. Each agent has its own terminal, with shared controls for supervising their work.

Aphelion is a desktop app for people who want character conversations, creative writing and coding help from models running on their own computer. It works offline after the first model download and doesn't require an account.

Memex is an AI journal for people who capture scattered notes, photos, and voice recordings and want help organizing them. It turns those fragments into structured records and connects information across entries.

Hatice is a self-hosted coding agent orchestrator for developers who want agents to work through issue-tracker tasks. It assigns issues to Claude Code agents and manages work across multiple turns.

Podcastfy turns source material into podcast-style audio conversations for creators and developers. It's an alternative to NotebookLM's podcast feature for people who want to customize the conversation or include podcast generation in their own software.