Coding Agents
Coding Agents don't just respond — they reason, plan, take action. They execute complex tasks autonomously and talk to other systems like they were built to. If you’re building something that works while you sleep, this is the layer that powers it.
-
WindsurfAgentic AI-native IDE (built on VS Code) from Cognition AI, powered by the Cascade agent for codebase-wide reasoning, multi-file edits, and autonomous terminal actions.
-
GitHub CopilotGitHub's AI pair programmer: context-aware code completions, chat, code review, and an autonomous coding agent built into VS Code, JetBrains, Visual Studio, and more.
-
PlandexOpen-source, terminal-based AI coding agent for large, multi-step tasks across many files. Handles up to 2M tokens of context, keeps generated changes in a review sandbox before applying them, and supports Claude, OpenAI, Gemini, and other providers. Now self-hosted only after Plandex Cloud wound down in October 2025.
-
Kilo CodeOpen-source agentic coding platform: a VS Code/JetBrains extension and CLI harness with five agent modes, 500+ model support at zero markup (BYOK), an MCP marketplace, and parallel subagent execution.
-
Cody (Sourcegraph)Sourcegraph's enterprise AI coding assistant, built on Code Graph for deep multi-repo context: code search, chat, and autocomplete tuned for large, complex codebases. Shifted to enterprise-only pricing in 2025.
-
Aider ChatOpen-source, git-aware AI pair programmer that runs entirely in your terminal. Maps your whole codebase, edits files, and auto-commits with sensible messages. Works with Claude, GPT-4o, DeepSeek, and more.
-
Blackbox AIMulti-agent coding assistant with a VS Code extension used across 35+ IDEs. Runs tasks in parallel across several AI models and picks the best result, with image-to-code conversion and an autonomous CyberCoder agent.
-
JetBrains AIJetBrains' built-in AI assistant: inline completion, chat, and the Junie autonomous agent, sharing one subscription and credit pool across IntelliJ-based IDEs. Junie plans, writes, tests, and self-corrects multi-file changes.
-
TabninePrivacy-focused AI code assistant with code completion, IDE chat, and agentic workflows. Supports SaaS, VPC, on-premises, and air-gapped deployment with zero code retention. Dropped its free plan in 2024.
-
Amazon Q Developer
AWS's AI coding assistant and successor to CodeWhisperer. Provides autonomous agents, security scanning, and code transformation, with console error diagnostics and AWS resource guidance built in.
-
Replit GhostwriterAI pair programmer built into the Replit cloud IDE, with inline completion, code explanation, natural-language refactors, and a chat assistant tuned for over 50 programming languages.
-
Google Code Assist (Gemini)
Google's AI coding assistant for VS Code, JetBrains, and Android Studio. Offers chat, completions, and agent mode with MCP support, grounded in BigQuery and Firebase. Dropped its free individual tier in June 2026 in favor of Antigravity.
-
Roo Code
VS Code extension that forked Cline to add customizable agent modes for different kinds of work, including a mode aimed at sysadmins using natural language in a CLI. Shut down on May 15, 2026, with its GitHub repository now archived.
-
Devin
Cognition AI's autonomous software engineer, which plans, writes, debugs, and ships code with limited supervision across OpenAI, Claude, Gemini, and Cognition's own SWE 1.6 models. Pricing runs from a free tier through Pro at $20/seat/month to Max at $200/seat/month, with usage billed in Agent Compute Units.
-
OpenHands
Open-source platform for autonomous coding agents, built by All Hands AI as a successor to OpenDevin. Give it a GitHub issue and it plans the fix, writes the code, runs tests, and opens a pull request inside a sandboxed Docker environment, with support for Claude, GPT-4o, Gemini, and open models through your own API key. MIT licensed, with around 90,000 GitHub stars.
-
Augment Code
Cloud AI coding platform for engineering teams, built around a Context Engine that indexes a whole codebase's call relationships and cross-service dependencies. Runs up to 10 autonomous Remote Agents in parallel, each in its own sandboxed VM, alongside IDE agents for VS Code and JetBrains. Indie plan is $20/month; Max is $200/month per developer.
-
ReSharperJetBrains' .NET productivity extension for Visual Studio, now adding Junie, an LLM-agnostic AI agent via the Agent Client Protocol that can refactor code, run terminal commands, manage git, and explain legacy code directly inside Visual Studio.
-
Cline
Open-source VS Code extension that turns the editor into a coding agent using your own API key for Claude, GPT, Gemini, or a local model. Reads and writes files, runs terminal commands, and browses the web for research inside a Plan/Act workflow, with an approval gate before each action. Over 5 million installs and 58,000+ GitHub stars.
-
Amazon Q Developer CLI
AWS's terminal agent, combining IDE-style autocomplete for CLIs like git and docker with natural-language chat and agentic execution. Reads and writes files locally, calls AWS APIs, runs bash commands, and can resolve merge conflicts with your permission.
-
Codename GooseOpen-source AI agent originally built at Block, runnable as a desktop app or CLI. Works with models from Anthropic, OpenAI, Google, Bedrock, OpenRouter, or local Ollama models, and connects to external systems via MCP. Now governed by the Agentic AI Foundation at the Linux Foundation.
-
Smol DeveloperLibrary for embedding a developer agent into your own app, letting it scaffold a full codebase from a single prompt. Started as a standalone CLI, then rewritten to be importable and smaller. Around 12,200 GitHub stars.
-
Rovo DevAtlassian's terminal coding agent, built into the CLI with over 300 commands for querying Jira, Bitbucket, and GitHub data through the Teamwork Graph. Completes and debugs code, writes tests, and generates documentation.
-
Gemini CLIGoogle's open-source terminal coding agent, built on Gemini models with a 1M-token context window. Shut down for free personal and Google AI Pro/Ultra accounts on June 18, 2026, replaced for consumer users by Antigravity 2.0; still usable with a Gemini API key, Vertex AI, or a Code Assist license.
-
DevOpsGPTOpen-source multi-agent system that turns natural-language requirements into working software, pulling and pushing code through Git. Works across any programming language and extends existing codebases, with around 6,000 GitHub stars.
-
PotpieOpen-source platform that turns a codebase, its history, tickets, and docs into a knowledge graph for AI agents. Ships pre-built agents for debugging, testing, code review, and Q&A that reason over that graph instead of a bare diff.
-
CosineAI software engineer that runs from the CLI, cloud, or desktop. Retrieves context from your codebase, plans and writes code changes, runs tests, and opens pull requests through its Genie agent, built on GPT-5.
-
Blinky Debugging AgentOpen-source AI debugging agent for VS Code, inspired by SWE-agent. Uses the VS Code API, Language Server Protocol, and print-statement debugging to track down and fix real backend bugs.
-
OpenAI CodexOpenAI's cloud-based autonomous coding agent, powered by the GPT-5 family. Reads a codebase, writes and edits code, runs tasks in a sandbox, and opens pull requests from a written instruction. Bundled into ChatGPT plans rather than sold separately.
-
AiderOpen-source, git-aware AI pair programmer that runs entirely in your terminal. Maps your whole codebase, edits files, and auto-commits with sensible messages. Works with Claude, GPT-4o, DeepSeek, and more. (Same project as "Aider Chat" in this directory, now maintained at Aider-AI/aider.)
-
VoidOpen-source, MIT-licensed code editor forked from VS Code, built as a transparent alternative to Cursor or Copilot. Supports inline editing, agentic task execution, autocomplete, and chat with local models via Ollama or APIs like OpenAI and Claude, with one-click import of existing VS Code extensions and themes.
-
CursorAI-first code editor (a VS Code fork) with Tab autocomplete, multi-file Agent mode, Composer, background agents, and native MCP support, running on frontier models like Claude, GPT, and Gemini.
-
AntigravityGoogle's agent-first development platform, built around Gemini 3. Antigravity 2.0 (announced at I/O 2026) adds a standalone desktop app, CLI, and SDK, parallel sub-agents that write components, configure routes, and run tests at the same time, and browser-in-the-loop testing where agents visually verify the UI they built. Free tier, with Pro at $20/month.
-
Claude CodeAnthropic's agentic coding tool, running in the terminal, IDE, desktop app, and browser. Reads your codebase, edits files, runs commands, and connects to your development tools. Bundled into every paid Claude plan starting at $20/month.
-
Trae AIByteDance's free AI-native code editor, a VS Code fork with chat, multi-file Agent mode, and a SOLO mode that drives larger tasks end to end. Includes MCP support and free access to Claude and GPT models.
-
StackSpot AIMulti-agent platform covering the full software development lifecycle. Lets teams orchestrate specialized agents for code generation, review, test writing, and deployment, with governance across GitHub, GitLab, and Bitbucket.
-
Refact.aiOpen-source AI coding agent that plans, executes, and iterates on engineering tasks end to end. Uses RAG for repository-aware completions, integrates with GitHub, databases, and CI/CD, and supports self-hosting with any major LLM.
-
ContinueOpen-source AI coding assistant for VS Code and JetBrains. Routes between Anthropic, OpenAI, Ollama, and other model providers through a single config file, with custom context providers, slash commands, and an Agent mode for multi-step tasks.
-
SweepOpen-source AI junior developer that turns GitHub issues directly into pull requests. Reads your codebase, plans the change, writes the code, and responds to review comments on its own PRs. Around 6,000 GitHub stars.
-
Kombai
AI agent built specifically for frontend development, turning Figma designs, images, or text instructions into production-ready UI code that fits your existing React, Tailwind, MUI, or Next.js stack. Reports a 96% compilation success rate.