Directory

See what a tool does and what you need to get started.

Search needs JavaScript. You can browse by category instead.

Coding agents · Showing 10 of 28

Filters
Clear all

Showing 1–10 of 10

Tools

A Claude Code plugin that coordinates work across multiple agents.

It distributes tasks among specialized agents and brings their results together through planning, execution and verification stages. It suits Claude Code users who want to divide coding work across a team of agents.

Needs Claude Code.

FreeTried

OpenAI's Codex Remote: start, steer and review coding tasks on connected development machines from the ChatGPT mobile app.

It connects Codex to your own development machines, so each task runs on the host and workspace you pick, and you follow it from your phone while away from the computer. It suits developers who work across several machines or want to keep an eye on an agent's work on the go.

Needs the ChatGPT mobile app and a Mac or Windows computer online

Free plan

ProtoMap by HoundDog.ai maps gRPC services, protobuf fields and their call sites, so coding agents see a change's impact without rediscovering it.

It analyzes gRPC services, methods, protobuf fields and call sites and gives that context to coding agents, so an agent knows the dependencies and the impact of a change before it edits code. It suits backend teams with many interconnected services.

Self-hosted: Docker Compose, 4 GB RAM, 20 GB disk

Free plan

Modern Web Guidance, from the Chrome, Edge, Firefox and MDN teams, gives coding agents browser-compatibility guidance and flags legacy code.

It's a set of guides and a scanner that give an agent browser-compatibility information and guidance on modern web features, and find legacy patterns in your code. It suits web developers who write code with coding agents and want output that fits today's browsers.

Free

The Code tab in the Claude desktop app runs coding sessions on your projects, with app previews, permission controls and diff review.

It runs Claude Code sessions on a project folder inside the Claude app itself: it previews the app as it's built, lets you set permission levels and shows the changes for review before you accept them. It suits people who want Claude Code with a visual interface instead of the terminal.

Paid

Anthropic's coding agent, in the terminal and in your editor.

It reads your whole project, edits files and runs commands to build a feature or fix a bug you describe in plain words. It suits developers, and anyone building software with AI who reviews what it changes.

Needs a Claude plan or an Anthropic API account

Paid

OpenAI's coding agent, in the terminal, the app and the cloud.

It writes and changes code in your project, either on your computer or in a cloud environment that keeps working on a task while you do something else. It suits developers, especially those who already use ChatGPT.

Needs a ChatGPT account or an OpenAI API key

Free plan

New in Cursor: cloud agents build and publish a web app for you, with hosting, an optional Postgres database and sign-in. Paid plans.

You ask for an app and get a working, hosted one with a link to share, without setting up any servers yourself. It suits teams that need quick internal tools; on individual plans you can also share the app publicly.

Needs a paid Cursor plan with cloud agents on

Paid

One window for Claude Code, Codex and other coding agents, on your computer or your phone.

It puts your coding agents in one interface, so you can switch between them and follow their work from your computer or your phone while they run on your own machine. It suits people who use more than one coding agent. It's free and open source.

Needs a supported coding agent you're signed in to

Free

Google's agent-first coding tool, with a free plan.

You describe a task and AI agents carry it out across your project folders, while you review each change. It suits developers who want several agents working at once instead of typing every edit themselves.

Needs a Google account

Free plan