Category · 30 tools
Agents, AI editors and prompt-to-app tools that turn what you describe into working code.
30 shown
An open-source command-line tool for AI pair programming with automatic git commits.
Terminal AI pairing with first-class git — every change is a commit you can undo. My pick when I want an agent that respects version control and runs on any model.
Visit ↗An AWS AI coding assistant with agentic chat, code generation and code transformation across popular IDEs.
Worth it mainly if you live in AWS — it knows your services and can do real code migrations. Outside that world, other assistants feel sharper.
Visit ↗A tool that turns Figma designs into React, Vue or HTML code.
Another Figma-to-code bridge, focused on developer handoff and live prototypes. Treat its output as scaffolding, not final code.
Visit ↗An agentic AI coding platform aimed at large, complex codebases via an IDE extension with a context engine.
Built for big, gnarly repos where context is everything. If your codebase is huge and other tools lose the thread, this is the one to trial.
Visit ↗In-browser AI that scaffolds, runs and deploys full-stack apps inside a WebContainer.
Great for spinning up a working prototype in one sitting without touching your local setup. Burns through tokens fast and complex apps outgrow it — but for demos and validation it is hard to beat.
Visit ↗A visual platform whose Visual Copilot converts Figma designs into React, Vue, Svelte and other framework code.
Its Visual Copilot turns Figma into real component code — the design-to-code bridge that actually works. Handy when handoff is the bottleneck.
Visit ↗Anthropic's terminal-native coding agent that reads your repo, edits files, runs commands and ships changes.
The most capable agent for real, multi-file work right now — it reasons about your whole repo instead of autocompleting the current line. Needs a paid Claude plan, but it earns it.
Deep dive →An open-source autonomous coding agent for VS Code and JetBrains with Plan/Act modes and MCP support.
Agentic power on your own API keys, with every action shown and approved before it runs. Great if you want autonomy without a subscription and don't mind paying for tokens.
Visit ↗Integration infrastructure that gives AI agents managed, authenticated access to 1,000+ third-party apps.
Managed, authenticated tool access for agents so you don't hand-roll OAuth for every integration. The plumbing layer for agents that actually do things.
Visit ↗An AI-first code editor (a VS Code fork) with inline edits, chat and agentic multi-file changes.
The most polished place to feel AI in your editor — tab-completion and inline edits are genuinely good. The free tier is enough to learn on before you hit the paywall.
Visit ↗Cognition's autonomous AI software engineer that completes coding tasks end to end in its own cloud environment.
An engineer you delegate whole tasks to, not an in-editor helper. Impressive for async work, but priced for teams and still needs a human watching.
Visit ↗Google's open-source AI agent that brings Gemini models into the terminal to query and edit codebases.
Google's open answer to Claude Code, with a generous free tier on a personal account. A no-cost way to get a capable agent in your terminal.
Visit ↗GitHub's AI assistant with completions, chat and an agent mode across major IDEs and GitHub itself.
The safe default that lives in every editor and ties straight into GitHub's PR flow. Less agentic than the newer crowd, but the free tier and ubiquity make it an easy first AI assistant.
Visit ↗AI features built into JetBrains IDEs, including the Junie autonomous agent, completion and context-aware chat.
If you live in IntelliJ or WebStorm, this with the Junie agent is the native fit. Outside JetBrains there's little reason to reach for it.
Visit ↗A tool that turns Figma and Adobe XD designs into front-end code for React, Next and more.
Design-to-code that converts your Figma into responsive front-end components. Useful for a head start on handoff; you'll still refactor the output.
Visit ↗A platform that turns prompts into deployed full-stack web apps with frontend, backend, database and auth.
Describe a full-stack app and get a deployed one with auth and a database wired up. The fastest prompt-to-product path — keep expectations realistic past the prototype.
Visit ↗An AI UI generator for product teams that creates interfaces and React code from prompts and screenshots.
Screenshot-to-UI and prompt-to-React with design-system constraints. A nice middle ground between v0 and a full app builder.
Visit ↗An open standard for connecting LLM applications to external data sources and tools.
Not a tool but the standard — the USB-C for plugging agents into your tools and data. Learn it; it's quietly becoming infrastructure for everything agentic.
Visit ↗An open-source AI-first visual editor for React apps that updates the underlying code as you edit.
Figma-like direct manipulation on top of a real React codebase, with edits written back to code. Promising for designers who want to touch the actual app.
Visit ↗An open-source platform for autonomous AI coding agents (formerly OpenDevin).
The leading open-source autonomous coding agent — delegate a task, watch it work in a sandbox. The free, self-hostable answer to Devin; bring your own model.
Visit ↗A browser dev platform whose Agent generates and edits full apps from natural-language prompts, with hosting built in.
Prompt an app into existence and it's already hosted with a database attached. Brilliant for idea-to-live-demo in the browser; you'll outgrow it for serious builds.
Visit ↗An AI assistant that uses code search to give context-aware answers across an organisation's repositories.
Code-search-grounded answers across an org's repos, now an enterprise product. Solid for big teams; solo devs get pointed elsewhere.
Visit ↗Google Labs' AI tool that generates UI designs and front-end code from prompts.
Google's free, AI-native design canvas — prompt a UI and get a design plus code. A genuine free contender in the prompt-to-UI space, evolving fast.
Visit ↗A code-first UI builder that designs with real React components and exports React and Tailwind CSS.
Design in real React components and export clean Tailwind — control without leaving the component model. Its MCP server lets agents use your design system too.
Visit ↗An AI code assistant focused on privacy, with SaaS, VPC, on-premises and air-gapped deployment options.
The enterprise-privacy pick — self-host or air-gap it and keep your code in-house. You pay for control, not for being the smartest model in the room.
Visit ↗A visual front-end builder that generates clean code, with AI and design import.
A visual builder that exports reasonably clean code across frameworks. A middle path between design tool and code editor.
Visit ↗An AI UI design tool that generates editable mockups from text prompts or screenshots.
Sketch or describe an interface and get an editable mockup — fast ideation for non-designers. Good for early concepts, not production design.
Visit ↗Vercel's prompt-to-UI generator that outputs React, Tailwind and shadcn/ui components.
Fastest way to go from a sentence to a decent-looking React component. Treat the output as a starting point, not a finished product — it leans generic if you don't push it.
Visit ↗An AI-native code editor with the Cascade agent for writing and editing code across a whole codebase.
A polished Cursor rival with a genuinely strong agent and a usable free tier. Worth a look — though its identity is in flux after the Cognition acquisition.
Visit ↗A high-performance native code editor written in Rust with built-in AI agent and edit prediction.
Blisteringly fast, with AI built in rather than bolted on. If Cursor's Electron weight bothers you, this is the performance answer — and it's open source.
Visit ↗