# Autonomous AI Coding Agents

> Source: https://theagentsindex.com/collections/autonomous-coding-agents (curated, quality-gated, re-verified)

An autonomous coding agent takes a task in plain English, plans the steps, edits the files, runs the tests, and reports back, usually inside its own sandbox and sometimes opening a pull request for human review. That is the bar this page draws the line at: a coding tool you can hand a non-trivial job to and walk away from, not a completion engine that needs a developer pressing keys at every step.

The 16 members below clear it on the published facts of their own listings; the methodology states the test so you can check ours against any tool you already use. Every entry is cross-checked against its listing's own `whatItDoes`, `keyFeatures` and `integrations` text, not the `autonomous` tag alone. That is why this set is shorter than the count of tagged listings and why a few near-misses (pair-programming tools, prompt-to-app builders, completion engines) are recorded as exclusions rather than left ambiguous.

Members are grouped by the surface the agent runs on: a terminal-native loop you drive from the shell, an agent inside an editor or IDE that already has the project open, or a managed cloud environment the vendor spins up for the run. The groupings describe how you interact with the agent, not how autonomous it is. Every member below meets the same bar in the methodology.

| Name | What it is | Why it's here | Pricing | API |
| --- | --- | --- | --- | --- |
| [Antigravity CLI](https://theagentsindex.com/antigravity-cli.md) | Google's terminal agent: a keyboard-driven TUI that edits code, spawns subagents and gates every tool call. | Google's terminal-native coding agent that replaced the open-source Gemini CLI in June 2026: same Antigravity 2.0 agent core that takes a task, plans the steps, edits the working tree and reports back, free for Individuals. | — | Yes |
| [Claude Code](https://theagentsindex.com/claude-code.md) | Agentic coding tool from Anthropic that edits your codebase, runs commands, and opens pull requests. | Anthropic's terminal-native coding agent with full-repository context: plans a multi-step task, edits files, runs commands, and reports results, included from the $17/month Claude Pro plan or billed per token through the API. | — | Yes |
| [Codex CLI](https://theagentsindex.com/codex-cli.md) | OpenAI's terminal coding agent for inspecting, editing and running repository code from the shell. | OpenAI's open-source terminal coding agent (Apache-2.0), bundled into ChatGPT plans from Free through Pro: plans, edits, runs commands and commits inside the working directory on a single task prompt. | — | Yes |
| [goose](https://theagentsindex.com/goose.md) | Open-source local AI agent — desktop app, CLI and ACP server — running on the LLM provider you bring. | Open-source AI coding agent (Apache 2.0) with a desktop app, CLI and 70+ MCP extensions across 15+ LLM providers, Linux Foundation project: runs a coding task end-to-end without human input, with the desktop app rendering interactive MCP UI inline. | — | Yes |
| [Muse Code](https://theagentsindex.com/muse-code.md) | Meta's terminal and CI coding agent, built on the Muse Spark model. | Meta's terminal coding agent that plans a coding task, executes it end-to-end and cross-checks against its own parallel runs: pay standard API rate, or opt into Contributor pricing for a steep discount in exchange for letting Meta train on your prompts. | — | Yes |
| [OpenCode](https://theagentsindex.com/opencode.md) | Open source AI coding agent for the terminal, desktop app and IDE, with 75+ model providers. | Free, open-source (MIT) AI coding agent for terminal, desktop and IDE with 75+ model providers and a tested model marketplace: takes a task, plans the changes, executes the edits, and reports back without developer steering. | — | Yes |
| [Cline](https://theagentsindex.com/cline.md) | Open-source coding agent for VS Code, JetBrains, the terminal and CI, running on whichever model you bring. | Open-source coding agent that runs in your editor, terminal or as an SDK with bring-your-own-key on 100+ models: plans a multi-file edit and executes it end-to-end without manual approval per step, with optional ClinePass for flat-rate open-model access. | — | Yes |
| [Cursor](https://theagentsindex.com/cursor.md) | Coding agent in a desktop editor, terminal CLI, cloud VMs and CI, priced per seat plus token usage. | Cursor's coding agent reads, plans and edits across your whole codebase with full project context: a free Hobby tier lets you try the autonomous flow before paying, and every paid tier ships Anthropic, OpenAI, Google and Cursor's own Grok 4.6. | — | Yes |
| [GitHub Copilot](https://theagentsindex.com/github-copilot.md) | GitHub's AI coding assistant and cloud agent, from editor completions to pull-request review. | GitHub's AI coding assistant with an issue-to-PR coding agent and a multi-agent control plane bundling Claude Code and OpenAI Codex: the agent reads an issue, plans the change, edits the code and opens a pull request, free tier with 2,000 completions/month. | — | Yes |
| [Junie](https://theagentsindex.com/junie.md) | JetBrains coding agent for the terminal, JetBrains IDEs and CI, running on hosted or your own model keys. | JetBrains' own AI coding agent: plans before it codes, drives the real IDE debugger, runs across the JetBrains IDE family and a standalone CLI on a single task prompt, free 30-day Pro trial then $10/month. | — | Yes |
| [Kilo Code](https://theagentsindex.com/kilo-code.md) | Open source AI coding agent for VS Code, JetBrains, CLI and cloud, with 500+ models at provider rates. | Open-source AI coding agent for VS Code, JetBrains, CLI and Cloud with 500+ models and no markup: takes a coding task and runs it end-to-end on a free Individual tier, with optional Kilo Pass credits for hosted model access. | — | Yes |
| [Augment Code](https://theagentsindex.com/augment-code.md) | Cosmos platform running fleets of coding agents across the SDLC, plus Auggie CLI and IDE extensions. | AI agent platform that automates code review, ticket-to-PR delivery, security remediation and incident response for dev teams: runs each task through to completion in its own environment, $100/month Business up to 50 seats. | — | Yes |
| [Devin](https://theagentsindex.com/devin.md) | Cloud AI software engineer that takes a ticket, works in its own VM, and opens a pull request. | Cognition's autonomous coding agent: takes a task through Slack, CLI or web, plans the change, edits, tests and opens a pull request inside its own sandbox, the canonical 2026 example of the autonomous-coding-agent shape. | — | Yes |
| [Factory](https://theagentsindex.com/factory-ai.md) | Droid coding agents that run in the desktop app, CLI, IDEs, CI and Factory-managed cloud computers. | Agent-native coding platform coordinating specialized Droids across the full software development lifecycle: from a $20/month Pro CLI that takes a task and runs it autonomously, to sales-gated enterprise deployment. | — | Yes |
| [OpenHands](https://theagentsindex.com/openhands.md) | Open-source software engineering agent with GUI, CLI, SDK and an enterprise control plane. | Open-source platform for cloud coding agents: the agent runs in a sandbox, plans the task, and edits files autonomously, free local OSS and a free Individual Cloud tier with 10 daily conversations. | — | Yes |
| [Replit Agent](https://theagentsindex.com/replit-agent.md) | Chat-driven agent that builds, tests and publishes full web and mobile apps on Replit's cloud. | AI agent inside Replit's browser IDE that builds, deploys and fixes full-stack apps from a chat prompt: runs autonomously across the build/test/fix loop until the task is done, free Starter plan with daily agent credits. | — | Yes |

## How this list is built

A tool earns a place here only if its own Published, sourced listing is in this site's coding-agents category AND describes a coding agent that takes a task, plans the steps, and executes them end-to-end without a human pressing keys at each step, which means reading, planning and editing the codebase itself, running the tests, and surfacing the result for review. A completion engine, a chat assistant that only suggests, or a tool that names a coding workflow but requires step-by-step human approval at every turn does not clear the bar. Verified directly against each listing's `whatItDoes`, `keyFeatures` and `integrations` text, never inferred from the `autonomous` tag alone. A tool that names a sandbox, opens a PR, or ships a 'task → result' loop in its own published copy clears the bar; a tool that needs the developer to copy-paste a suggestion into the editor does not.