AI Apps

A browser agent from Browser Use and TypeSafe: give it a goal and it clicks, types and scrolls through Chrome on its own.

• Workflow automation
• Agent framework

Baidu's open-source OCR models and inference code for parsing long, multi-page images and PDFs. You deploy it on your own GPU.

• PDF processing
• Data and content transformation
• +1

xAI's official terminal-based AI coding agent: it reads code, edits files, runs commands, and can run headless in CI.

• AI coding
• Software debugging
• +1

Karpathy's autoresearch: an AI agent spends the night on one GPU editing code, training a small model and keeping the changes that work.

• Model training
• Research assistance

Turns codebases and personal sources into a Markdown wiki that keeps itself updated and that coding agents can query.

• Knowledge management
• Code quality and comprehension
• +2

A multilingual decision model that answers an agent's multiple-choice questions with options and probabilities in one pass.

• Agent decision model
• Tool and API integration

A pure C local inference engine that runs very large open models on an ordinary computer's CPU.

• Local model runtime
• Model inference and serving
• +1

DeepSeek's official harness for coding agents: one npx command opens a local web interface, and it can load plugins.

• AI coding
• Agent framework
• +1

A self-hosted AI gateway that puts several model providers' accounts, keys and quotas behind one interface.

• Tool and API integration
• App and service management
• +1

A local gateway that puts your AI coding agents' models, subscriptions and accounts in one place, switchable from the menu bar.

• App and service management
• Tool and API integration

A headless browser engine written in Rust, made for AI agents to open pages, click and read content.

• Tool and API integration
• Workflow automation

Connects your AI coding agent to reverse-engineering tools for inspecting programs, binaries and websites without source code.

• Code quality and comprehension
• Tool and API integration

A macOS menu-bar tool that shows the status and usage of AI coding agents like Claude and Codex in a Dynamic Island.

• Agent monitoring
• Usage and cost tracking

Compresses logs, files and conversation history on your machine before they reach the LLM, so coding agents use fewer tokens.

• Context management

An open-source AI coding agent that reads, edits and runs your project from the terminal or a desktop app, using the model you choose.

• AI coding

An open-source foundation model for financial candlestick (K-line) data that forecasts price movements, for research use, not investment advice.

• Financial model research

A desktop workbench for running several coding agents side by side, each in its own worktree, then comparing results and merging the best.

• Multi-agent orchestration
• AI coding

Generates command-line interfaces for desktop software so AI agents can drive tools like GIMP, Blender and LibreOffice.

• Tool and API integration
• Workflow automation

A local search engine that indexes your code, docs and logs so a coding agent can search them instead of reading files one by one.

• Code search
• AI search

An open-source engine for running and serving large language models, turning Hugging Face models into an API you can call.

• Model inference and serving
• Developer deployment

OpenAI's official command-line tool for generating images, uploading files, submitting batches and checking API results from the terminal.

• Tool and API integration
• Image generation
• +1

A DeepSeek Harness plugin that lays out the context, tool calls and attachments inside each AI coding agent session.

• Context management
• Agent monitoring

DeepSeek's open-source high-performance inference framework, the one behind its own large-scale production services, for GPUs and NPUs.

• Model inference and serving
• Developer deployment

A guide and template kit for building with AI coding agents, with three paths: set up your machine, start a new app, or take over an existing one.

• AI coding

A Google Research framework that improves an agent's prompts, tools, memory and skills while it works, without changing the model itself.

• Agent framework
• Model evaluation
• +1

An open-source app for running large models on your own computer to chat, code and read images, and to serve local models to Claude Code and Codex.

• Local model runtime
• General AI assistant
• +1

An experiment that rewrites ordinary JavaScript into a neural network, with no training and exactly the same results.

• Model training

An open-source tool that makes AI coding agents answer in fewer words; we recommend its caveman Skill, which leads with the answer and keeps code and errors exactly as they are.

• AI coding
• Context management

An open-source, self-hostable observability platform that puts logs, traces, errors, session replays and AI call tracing in one place.

• Agent monitoring
• Application observability

Open-source text-to-speech models from Resemble AI: give them a short reference clip and they read your text in that voice, in 23 languages including Chinese.

• Text-to-speech
• Voice cloning

A companion that lives in your Mac's notch, shows what your coding agents are doing and lets you approve or deny their permission requests on the spot.

• Agent monitoring
• AI coding