deepseek-ai/deepseek-harness
DeepSeek Harness (dsh) is an open-source agent harness from DeepSeek AI built on an everything-is-a-plugin architecture over the Cordis framework, started from npm with a local web UI.
Awesome AI Agents › AI Developer Tools
DeepSeek Harness (dsh) is an open-source agent harness from DeepSeek AI built on an everything-is-a-plugin architecture over the Cordis framework, started from npm with a local web UI.
Open-source terminal coding agent written in Rust that edits files, runs commands and works toward a goal across many model providers.
Tabby is a self-hosted, open-source AI coding assistant providing a private and customizable alternative to GitHub Copilot with rich integration and deployment options.
Cursor is an AI-powered code editor designed to enhance programming productivity by integrating intelligent code suggestions and AI-driven features directly into the coding workflow.
Serena is a free and open-source coding agent toolkit that enhances large language models with semantic code retrieval and editing capabilities, enabling autonomous coding workflows directly on your codebase.
Open-source terminal coding agent from the Qwen team with subagents, skills, hooks and MCP support, speaking the OpenAI, Anthropic, Gemini and Qwen API formats plus local model runtimes.
Plandex is an open-source AI coding agent designed to plan, execute, and manage large-scale coding tasks and real-world projects with high efficiency and flexibility.
Smol Developer is a library and toolset that embeds a junior developer AI agent into applications to scaffold and generate codebases interactively through human-in-the-loop prompt engineering.
Open-source workspace for running, supervising and reviewing AI coding agents from desktop, browser, VS Code and mobile, with session goals, parallel multi-model runs and diff walkthroughs.
Hunk is a review-first terminal diff viewer for agent-authored changesets, built on OpenTUI and Pierre diffs, with inline agent annotations, watch mode and Git, Jujutsu and Sapling support.
Recaf is a modern, user-friendly Java bytecode editor that simplifies bytecode manipulation, supports decompilation, recompilation, deobfuscation, and extends functionality through scripting and plugins.
Awesome Vibe Coding is a curated list of resources and tools that support the innovative practice of vibe coding, where developers collaborate with AI to write and deploy code more intuitively and efficiently.
ByteRover CLI (brv) provides persistent, structured memory for AI coding agents, allowing developers to curate context, share knowledge, and interact with LLMs.
Markdown-defined coding agents on top of the OpenCode CLI that load project-specific patterns before generating code and require human approval before every execution step.
AutoDev is an AI-powered coding assistant plugin for IntelliJ IDEA that offers multilingual support, automatic code generation, bug detection, customizable prompts, and a suite of interactive development tools to enhance the software development lifecycle.
CLI and MCP server that bundles a prompt with selected project files, sends the context to one or more models via API or a signed-in browser session, and stores each run as a replayable session.
Terminal coding agent distributed as an npm CLI that ships features, fixes bugs and refactors code while learning a personal coding taste profile from every accepted, rejected and edited change.
Refact is an open-source AI software development agent that automates end-to-end engineering tasks by integrating with developer tools, supporting multiple programming languages, and providing intelligent code assistance and automation.
Devon is an open-source AI-powered pair programming tool that assists developers with multi-file editing, code exploration, bug fixing, and test writing through integration with advanced AI models and a user-friendly interface.
Open-source terminal coding agent from the Nano Collective that runs on local Ollama models or any OpenAI-compatible API, with skills, subagents, MCP servers, checkpointing and plan or yolo modes.
RA.Aid is an autonomous software development agent that uses a three-stage AI-driven process to research, plan, and implement complex multi-step programming tasks with optional human interaction and advanced AI integrations.
Pilot Shell, an installable workflow layer for Claude Code and Codex CLI adding spec-driven commands, enforced quality gates, persistent memory, code intelligence and a local console.
Desktop and browser application from Coder for running several coding agents in parallel across isolated local, git worktree or SSH workspaces, with git divergence views, multi-model support and cost tracking.
Rust CLI agent for DevOps work that runs continuously in the background, generates infrastructure code and debugs deployments while keeping credentials hidden from the model.
ReactAgent is an open-source autonomous agent that uses GPT-4 to generate and compose React components from user stories, facilitating rapid UI prototyping and development.
Open-source CLI that adds a spec-driven, repo-native workflow around AI coding agents, storing specs, plans, work packages and review state in Git and running agents in isolated worktrees.
Reference implementation of the AIDE algorithm: an LLM agent that performs tree search over machine learning code, drafting, debugging and benchmarking scripts until a user-defined metric is optimised.
Open-source CLI coding assistant that routes each step of a task to a different model tier and chains multi-step workflows, called superpowers, for review, refactoring, debugging, explanation and test generation.
Cougar CLI is a TypeScript command-line programming assistant that talks to OpenRouter, OpenAI, Claude, Zhipu or any OpenAI-compatible provider, with image input support on Windows.
Command-line agent that turns a website into a typed API client: a browser captures network traffic to a HAR file and a model writes a working client for the endpoints the site actually uses.
OpenContext is a personal context store for coding agents, pairing an oc CLI, an MCP server, generated skills and slash commands, and a desktop GUI over a global library of Markdown knowledge.
An installable skill for coding agents that restructures a task into a subagent workflow with planning, testing and review, reported to solve 73 of 89 Terminal-Bench 2.1 tasks against 60 for the bare agent.
Telegram bridge for the Codex, Claude Code, OpenCode and pi coding agents that streams progress, manages projects and git worktrees, and resumes sessions from a phone or a terminal.
Agent Deck is a Go terminal session manager for AI coding agents that shows every Claude Code, Codex, Gemini or OpenCode session in one TUI with groups, search, forking, git worktrees and cost tracking.
CLI orchestrator that runs multiple coding agents as autonomous workers: it pulls issues from Linear or a local tracker, runs worker and reviewer pipelines, reports to chat, and keeps per-repository memory in LanceDB.
Cyrus is a self-hostable background coding agent that watches Linear, GitHub, GitLab and Slack issues, runs Claude Code, Codex, Cursor or Gemini in isolated git worktrees and streams progress back.
OpenCode Portal is a mobile-first web interface for the OpenCode AI coding agent, giving browser access to sessions, chat, file mentions and model selection on a remote machine.
SWE-smith is a toolkit for creating task instances, generating solution trajectories, and training software engineering agents to improve their capabilities using scalable data from Python repositories.
Beads UI is a local-first web interface for the bd issue-tracker CLI, giving issues, epics and board views so developers can review and edit work items shared with a coding agent.
hof is a Go CLI built on CUE that unifies data modelling, schemas, deterministic and agentic code generation and a DAG task engine, with agent servers, an LLM chat mode and a terminal interface.
Autopilot is an AI-driven tool that uses GPT to read codebases, create context, and automatically solve programming tasks by modifying relevant files.
claude-sub-agent is a set of Claude Code subagent definitions and a slash command that run a phased software development pipeline, moving a project idea through requirements, architecture, planning, implementation, testing and review with quality gates between phases.
Self-hosted Go bot that runs a team of Claude Code subagents on your own server and drives it from Telegram or VK: describe a feature in chat and get a branch, tests, a review and a pull request.
TalkCody is an open-source desktop coding agent built with Tauri, Rust and React that runs tasks at four levels of parallelism, stores everything locally, and works with hosted or local models through MCP servers and shared skills.
Kolosal CLI is a terminal coding agent based on Qwen Code that reads and edits large codebases, automates git and review chores, and runs against OpenAI-compatible or Hugging Face model endpoints.
Live-SWE-agent is a self-evolving software engineering agent scaffold built on mini-swe-agent that extends its own tooling at runtime, with a public SWE-bench Verified and SWE-Bench Pro leaderboard.
Micro Agent is a minimalistic autonomous agent powered by OpenAI GPT-4 that autonomously writes and tests Python software based on a given purpose, serving as a research tool towards AGI and AI-driven programming.
Zero-config coding agent shipped as a single static Go binary, where a planning orchestrator spawns ephemeral subagents that absorb tool noise and are destroyed, keeping the main context small.
The English Compiler is a proof-of-concept AI tool that converts English markdown specifications directly into functional code, demonstrating a future where natural language can replace traditional programming.
bumpgen is an AI-powered tool that automates upgrading npm packages and fixes breaking changes in TypeScript and TSX projects.
LibreCode is a free, open-source, AI-powered code editor built with .NET 10 and Avalonia UI, offering local AI assistance, reverse engineering toolkits for .NET, WebAssembly, and ELF binaries, and ...