Coding Agents

One intelligence. One workspace. One place for everything.

OYYO ONE AI is an intelligent AI workspace that turns human intent into execution. OYYO Intelligence understands the objective, context, files, memory and available capabilities, then coordinates the right AI models, agents, tools and connectors for the task. It combines native OYYO models, external cloud AI, local AI, persistent memory, Prompt Knowledge, files, APIs, browser automation and workflows in one environment. Users can research, code, analyze data, create documents, spreadsheets and PDFs, automate work and collaborate without constantly switching tools. One intelligence. One workspace. One place for everything. Built in Dubai. Made in UAE.

The AI software engineer that ships code autonomously

Devin by Cognition is an autonomous AI software engineer that plans tasks, writes code, runs tests and opens pull requests with minimal supervision.

MathGPT is a free AI math solver that shows every step.

MathGPT is a free AI math solver that helps users solve mathematical problems step-by-step. Users can type a problem, upload a photo, or use the MathGPT math keyboard to input their queries. The platform specializes in a range of topics, including algebra, geometry, trigonometry, statistics, and calculus. Unique features of MathGPT include transcribing handwritten problems, providing detailed solutions, and offering practice questions without requiring any sign-up or payment.

Anthropic's agentic coding tool that lives in your terminal

Claude Code is an agentic coding assistant that understands your codebase, edits files, runs commands and completes entire engineering tasks from the terminal or IDE.

A free desktop AI assistant that runs open models on your computer, with no account.

Draggy is a free, open-source desktop AI assistant for Windows, macOS, and Linux. It runs open models on your computer and can also connect to local servers such as Ollama and LM Studio, or to cloud providers you add. It can answer questions using web pages and sources, edit code in folders, work by voice, and open pages in its built-in browser. Chats and files stay on your disk by default; the page says conversations leave your computer only if you add a cloud provider.

The AI code editor built to make you extraordinarily productive

Cursor is an AI-first code editor with an autonomous agent mode that can plan, edit and verify multi-file changes across your codebase.

Open-source agent that turns labeled GitHub issues into reviewed, merged PRs and tagged releases

Orbi is an open-source (AGPL-3.0) AI coding agent that works from GitHub issues. Label an issue ai-ready and it writes the change in an isolated worktree, opens a pull request, has a separate review session check it against the issue's acceptance criteria, merges only what passed review, and cuts a tagged release. It also handles ops tickets such as deploys or investigations, posting every real command and its output on the issue. Self-host it free with your own model key, or use Orbi Cloud with the first 3 merged deliveries free.

Bring an AI agent into your application.

CornAgent is an open-source full-stack AI agent designed for flexibility and ease of embedding. Built with React and FastAPI, it allows users to self-host and enjoy advanced features such as streaming, pause and resume functionality, durable state management, and support for parallel multi-agent subtasks.

A public birth certificate for your AI agent — provable without us.

DEUSPROOF is an open registry of AI agents. Every agent gets a public birth certificate (an inscription number, a date and a face drawn from its identifier) and can take possession of it with a key it generates itself. Agents can also timestamp what they make by sending only a SHA-256, check prior art before publishing, and verify any record: RFC 3161 timestamps and Bitcoin anchors anyone can check without trusting us. It proves when, not who. Free, no account. Remote MCP at https://deusproof.com/mcp, plus A2A 1.0 and a plain HTTP API.

Your AI agents, governed.

AgentPlaza is an open-source platform for running AI agents on your own hardware, from one machine or across a fleet. It provides a console for overseeing agent plans, conversations, tasks and releases. Each agent runs in its own microVM and is governed by laws that the owner can change. The platform includes an orchestrator, planning and review workflows, a treasury for task payments, and controls for tools, secrets, snapshots and agent approvals.

From App Idea to Mobile App in Minutes.

Gen AI X App is an AI app builder for Flutter that transforms text prompts into native iOS and Android mobile applications. With capabilities like live cloud compilation, QR code testing on physical devices, and GitHub integration, it aims to streamline the app development process for users. The platform is designed for anyone looking to create mobile applications quickly and efficiently using natural language inputs.

Code with Claude or ChatGPT from your terminal.

agent is a native terminal coding agent for macOS and Linux that works with Claude or ChatGPT. It installs with one command and updates itself.

AI coding for the next generation of developers

GitWit is an open source, AI-native coding platform that allows you to build your ideas in minutes. It offers a powerful browser-based IDE along with AI-powered assistance for developers, facilitating faster coding, testing, and deployment. The platform encourages community participation and is committed to shaping the future of AI development.

Synthetic data that feels real

Generate and reuse names, avatars, identities, products and company rosters.

Your launch and growth copilot

Actuant runs a launch-readiness audit on your deployed app in a real browser. 52 deterministic checks scored 0 to 100 across security, SEO, AI-search visibility, performance and conversion, then opens the fixes as pull requests you review and re-audits after every deploy.

Free, open-source desktop and web app for Meta's Muse Code CLI.

Helicon is a free, MIT-licensed desktop app (Windows, macOS, Linux) and VS Code extension for Meta's Muse Code CLI, allowing users to manage their projects, sessions, and approvals all in one window. It integrates seamlessly with existing Muse Code subscriptions, eliminating the need for additional billing. Helicon empowers users with tools to efficiently handle coding tasks without the clutter of terminal commands, providing a user-friendly interface for both beginners and experienced developers. Helicon is unofficial and not affiliated with Meta.

Easily create aesthetic, animated architecture diagrams with archgram.

archgram is an open-source CLI tool for generating software architecture diagrams. It allows users to write a YAML or JSON specification, or let AI agents like Claude Code generate diagrams directly from code. The output is a self-contained SVG that adapts to both light and dark themes and animates the flow of requests. This makes it ideal for integrating system architecture visuals directly into documentation such as GitHub READMEs.

The open-source notebook your AI tools write to.

LockIn is an open-source notebook designed to save notes generated by your AI tools. It allows users to effectively manage and store insights from models like Claude, Codex, and Cursor. The application connects agents via the Model Context Protocol (MCP) and ensures that your notes are organized by subject, making them easy to retrieve later.

Open source AI code review that runs before you push.

Qodex.ai offers an open source AI code review tool called OpenQodex. This tool is designed for developers who need to perform code reviews before pushing changes, utilizing AI models like Claude Code and Codex. OpenQodex analyzes changes by running multiple scanners on the code, ensuring every finding is checked by a separate reviewer, resulting in a thorough report.

Coding agent for open models

Open Interpreter is an open-source coding agent designed for low-cost and open-weight models. It enables users to execute commands, read files, and edit code through a terminal interface, encapsulating its operations within a secured environment. The tool aims to streamline coding tasks by integrating various models optimally suited for specific functionalities.

Your team’s Claude Code, Codex, Cursor and Gemini CLI agents in one shared project — memory, tasks,

Joxo is the team channel for AI coding agents, providing a simple way to connect with teammates without requiring an enterprise account. Users can seamlessly integrate their existing AI agent subscriptions, allowing collaborative work on shared projects with a unified task board and memory. It streamlines communication between coding agents, enhancing productivity and project flow.

Review a CLAUDE.md, an AGENTS.md or one rule, then copy an evidence-based refactoring prompt for you

Disregard is a tool designed for reviewing CLAUDE.md and AGENTS.md files. Users can paste or upload their markdown files and receive a prompt that helps refine AI agent instructions. By analyzing these files, the site assists in optimizing rules and improving the functionality of AI agents. It is particularly valuable for anyone working with AI models and looking to enhance their command inputs.

A local video studio for your coding agent.

showtime is a local video studio designed specifically for your coding agent. It enables users to create videos through motion graphics, voice-over, music, captions, and footage editing all from a simple sentence. The service operates without relying on cloud AI services, API keys, or uploads, ensuring a seamless local rendering experience.

Native SSH client for Apple Silicon Mac: tabs, splits, SFTP, and host monitoring.

Docklark is a native SSH client for Apple Silicon Mac. It puts connections, terminal tabs, split panes, SFTP, and host monitoring in one workspace, in English and Simplified Chinese. It is a desktop Mac app for people who work in the terminal and want sessions, files, and host status together.

Run an agent that reads and edits files on your server, and continue the conversation from any devic

Pipaya is an open-source AI coding agent that runs on your server. It can read and edit files and run commands, with command execution requiring approval.

Forgot the command? Just ask.

lxask is an offline Linux command assistant that turns plain-language requests into commands, explains each part, and checks commands against man pages and tldr pages. It runs locally on your machine, needs no API key or cloud connection, and does not run commands for you. It is for Linux users who want help composing commands. It supports restrictions in requests, asks follow-up questions when a request is vague, and can put a suggested command on the prompt for editing.

Decompile and analyze native binaries, Android APKs, and malicious npm, PyPI, PowerShell & shell pac

OpenBin.ai is a free, open-source reverse engineering platform for researchers analyzing native binaries, Android APKs, npm and PyPI packages, PowerShell, and shell scripts. Its workspace provides decompiled code analysis, hotspots, call chains, string and import analysis, and pseudo-C with citations. Researchers can use their own LLM provider key, collaborate through published reports and citations, and access projects from anywhere. The CLI runs Ghidra locally, so binaries do not leave the user's machine for local decompilation. BINNY, a beta terminal agent, can investigate samples and draft reports.

A fullscreen TUI that pairs a real, interactive shell with an AI agent chat.

Sensus is a terminal-based user interface that integrates an interactive AI agent with a traditional shell. The AI agent can observe the terminal, perform tasks in the background, and assist users in real-time. It is designed for developers and system administrators seeking a modern and efficient command-line experience.

Shape software together.

Flowfield is a workspace designed for collaborative software development, enabling developers to work seamlessly with AI. It allows teams to plan projects with the Coordinator, build using parallel agents, and refine their work on a visual task board. This innovative approach fosters creativity and efficiency in tackling ambitious projects.

A free skill library for AI agents.

The Playbook is a free library of skills for AI agents, from Muse FM. Each skill is a tested, step-by-step playbook an agent can read and follow to do a real job — no guessing, no hallucinated workflows. Free to use; an optional paid tier (Playbook Pro) adds skill bundles, reports, and intel feeds.

see what a file would make your AI agent do

Sentinel is an AI tool designed to analyze project files and reveal hidden instructions that AI coding agents would follow. It reads files such as CLAUDE.md and AGENTS.md to inform users about potentially dangerous commands that could compromise security. This tool is valuable for developers looking to ensure their projects remain safe from unintended consequences of automated scripts.

Spawn and control Oh My Pi (OMP) tasks on your own computers and servers from Android.

PinkCollab provides a bridge to Oh My Pi (OMP), enabling users to spawn and control tasks directly from their Android devices. The platform allows for seamless interaction with OMP processes hosted on users' machines, promoting efficient task management without cloud intermediaries. Users can engage with their tasks in real-time, providing inputs and monitoring progress all from a mobile interface.

A control plane for coding agents that you own: Codex, Claude Code, Cursor, Gemini and any ACP agent

HarnessDesk is a desktop application designed for developers managing multiple coding agents. It offers a central control plane where you can oversee models like Codex, Claude Code, and Cursor, along with tracking context, policy, history, and costs in one unified dashboard. This tool operates with various agents and supports seamless hand-offs, ensuring that your workflow remains uninterrupted even when agents run out of resources.

The visual app planner for anyone building with AI.

Flowmaps is a visual app planner for anyone building with AI. Describe your idea and it becomes a mind map of every feature your app needs, with what to build first. It flags gaps like security, payments and privacy, then writes step-by-step build prompts for your AI builder.

EU inference API for agents: OpenAI- and Anthropic-compatible, zero data retention on own models

Heabsy AI Inference is an OpenAI- and Anthropic-compatible API from an EU company (FEYA, s.r.o., Bratislava). The flagship model, Qwen3.8 27B, runs on GPUs we operate in the EEA; prompts and completions are processed in memory only, not stored and not used for training. Our other own open models, such as MiMo, also run on our GPUs in the EEA with zero data retention, and partner-hosted models are labelled in the catalog. Keys can be limited to the EEA. Agents get streaming, tool calling, structured output and prefix caching; one key, one DPA and one EU invoice. Per-token pricing at heabsy.com/models.

The Local-First Autonomous Desktop AI Agent

Kyro is an open-source desktop AI agent for developers and knowledge workers. It combines a Rust native core, Tauri v2, and React 19, and runs locally with SQLite storage and no cloud telemetry. Kyro supports desktop automation, document analysis, browser automation, and developer workflows. Its features include a reversible file organizer, a screen explainer, managed development servers, Git checkpoints, and specialist coding agents running in isolated worktrees. It uses a centralized authorization engine and fail-closed security controls.