The vibecoding toolbox

A VS Code-based editor rebuilt around an in-line agent, Tab autocomplete, and full-codebase chat. Its Agent can plan multi-file edits, run terminal commands, and iterate on failures while you stay in the loop. The most widely adopted AI IDE in the scene.

Pricing Free Hobby tier; Pro $20/mo, Pro+ $60/mo, Ultra $200/mo, Teams $40/user/mo

Best for Anyone who wants the best-supported, most popular AI editor out of the box

Formerly Codeium, Windsurf is an AI-native editor built around Cascade — an agent that keeps your intent, edits, and terminal in sync. After Cognition (makers of Devin) acquired it, it added the fast in-house SWE-1.5 model and Codemaps for visualizing how a change ripples through a codebase.

Pricing Free tier with quotas; Pro $20/mo, Teams $30/user/mo, Max $200/mo

Best for Developers who want a polished flow-state agent and tight codebase awareness

An open-source, GPU-accelerated editor written in Rust, with AI woven through every buffer. Its Agent Panel runs multiple agents in parallel via the open Agent Client Protocol (ACP), so you can drive Claude, Codex, and others against different parts of the codebase at once. Bring your own keys or use Zed-hosted models.

Pricing Editor is free & open source; optional hosted AI plans, or bring your own API keys

Best for Performance-obsessed devs who want speed, openness, and parallel agents

An agentic IDE from AWS that turns a prompt into a formal spec, then generates and maintains code against it — making the spec, not the diff, the unit of work. Agent Hooks watch your filesystem and trigger AI actions on change. A VS Code fork with a deep agent layer, powered by models on Amazon Bedrock.

Pricing Free during preview; announced GA tiers Pro $19/user/mo and Pro+ $39/user/mo

Best for Teams who want engineering rigor and durable specs, not just throwaway prompts

A terminal-native coding agent from Anthropic that reads your repo, edits files, runs commands, and manages git directly from the shell. It supports subagents, custom skills, and MCP servers, and plugs into editors and CI. Billed through your Claude subscription or the API.

Pricing Via Claude Pro ($20/mo) or Max ($100–$200/mo), or pay-as-you-go API; no free tier

Best for Terminal-first developers who want a deeply capable agent on real codebases

OpenAI’s open-source terminal agent that reads, writes, and runs code locally or in the cloud, with a sandbox and approval modes for safety. It pairs with the Codex cloud workspace and IDE extension, sharing one account across surfaces. Included with ChatGPT paid plans rather than sold separately.

Pricing Bundled with ChatGPT Plus ($20/mo) and up; token-based usage, or pay-as-you-go API

Best for ChatGPT subscribers who want an agent that spans terminal, cloud, and editor

An open-source terminal pair-programmer that maps your whole repo, edits across files, and auto-commits each change with a sensible message so every step is easy to diff or undo. Its architect/editor mode pairs a strong reasoning model for planning with a fast model for edits. Model-agnostic — point it at almost any LLM, including local ones.

Pricing Open source & free; you pay your chosen model provider (or run a local model)

Best for Git-disciplined devs who want a model-agnostic, scriptable terminal partner

A fast, themeable terminal coding agent from the SST team, with a native TUI, automatic LSP loading, and 75+ model providers via Models.dev. Run multiple sessions in parallel, switch between full-access "build" and read-only "plan" agents, and share session links for review. Log in with a Claude plan or bring any key.

Pricing Open source & free; bring your own model keys or a Claude/other subscription

Best for Terminal lovers who want an open, provider-agnostic alternative to closed agents

StackBlitz’s in-browser agent that builds, runs, edits, and deploys full-stack web apps from a single prompt — no local setup. It runs Node and npm entirely in the browser via WebContainers, so you see a live preview as it works. Billed by AI token usage.

Pricing Free daily token allowance; Pro from $20/mo, Teams $30/user/mo

Best for Rapid prototyping and shipping web apps without touching a local toolchain

A conversational app builder that generates frontend, backend, auth, and database from prompts plus direct visual editing. Native Supabase and Stripe integrations mean real data, logins, and payments — not just a mockup. Popular with founders shipping MVPs fast.

Pricing Free daily credits; Pro $25/mo, Business $50/mo (credit-based)

Best for Non-engineers and founders building real, data-backed products by chatting

An autonomous agent inside Replit’s cloud workspace that builds full-stack apps from natural language, then debugs, writes tests, manages databases, and deploys — all without leaving the browser. The current Agent 3 runs long autonomous sessions. Effort-based pricing means simple changes cost less than complex ones.

Pricing Free with limited agent access; Core $25/mo (effort-based credits), Teams $40/user/mo

Best for People who want one tool that builds, hosts, and runs the whole app

Vercel’s generative-UI tool (now v0.app) that turns prompts into production-ready React with shadcn/ui and Tailwind. In 2026 it grew into a full Next.js sandbox with API routes, server actions, Supabase, a git panel for branches and PRs, and one-click deploy to Vercel.

Pricing Free tier with monthly credits; Premium $20/mo, Team $30/user/mo (token-based)

Best for React and Next.js teams who want polished UI and shippable git workflows

A browser-based, agentic dev environment from Google that unifies Project IDX with Gemini-powered agents. Its App Prototyping agent turns natural-language, sketches, or screenshots into a working full-stack web app you can refine, test, and share — with the whole Firebase backend a click away.

Pricing Free preview with a few workspaces (more via Google Developer Program); paid tiers expand quota

Best for Builders already in Google/Firebase land who want prototyping plus a real backend

GitHub’s assistant for VS Code, JetBrains, and the GitHub site, offering completions, chat, code review, and a coding agent. Agent mode plans and executes multi-step tasks across files and the terminal, and you can pick among models. Deeply integrated with pull requests and the GitHub workflow.

Pricing Free tier (limited); Pro $10/mo, Pro+ $39/mo, Business $19/user/mo, Enterprise $39/user/mo

Best for Teams living in GitHub who want a mature, broadly-integrated assistant

An Apache-2.0 VS Code extension (formerly Claude Dev) that reads and edits files, runs terminal commands with approval, and works through tasks with human checkpoints. It is bring-your-own-key — connect Anthropic, OpenAI, OpenRouter, Bedrock, Vertex, or local models and pay providers directly. Also ships as an SDK and CLI.

Pricing Free & open source; bring your own API key and pay the model provider directly

Best for Devs who want a transparent, open agent inside VS Code with full key control

A fast-moving open-source fork of Cline built around custom modes — Code, Architect, Ask, Debug, and personas you define with scoped tool permissions. "Boomerang" sub-tasks return results to the parent thread and a memory bank persists project context. Diff-based edits keep token costs down.

Pricing Free & open source; bring your own API key and pay the model provider directly

Best for Power users who want granular modes, custom personas, and rapid releases

An MIT-licensed assistant for VS Code, JetBrains, and Neovim with chat, autocomplete, and an agent mode that plans and executes multi-step tasks. In 2026 it leaned into source-controlled AI rules and continuous AI — checks you can enforce in CI to keep a "software factory" consistent. Fully bring-your-own-model.

Pricing Free & open source; bring your own model (cloud, local, or 100+ providers)

Best for Privacy-minded or JetBrains devs who want open, customizable, CI-enforceable AI

An enterprise-grade assistant built for compliance: zero code retention, no training on your code, and SaaS, VPC, on-prem, or fully air-gapped deployment. It offers completions, SDLC chat, and an agentic platform across major IDEs, with access to multiple LLMs. SOC 2, GDPR, and ISO 27001 certified.

Pricing Code Assistant $39/user/mo, Agentic Platform $59/user/mo, Enterprise custom (14-day trial)

Best for Regulated teams (finance, healthcare, defense) that need code to stay in-house

An agent for VS Code and JetBrains whose Context Engine indexes hundreds of thousands of files across many repos, so it understands how a change connects across a large system. It runs MCP and native tools, does not train on your data, and is SOC 2 Type II. The standout pick when scale is the problem.

Pricing Indie $20/mo, Standard $60/mo, Max $200/mo (credit-based, with auto top-up)

Best for Engineers in large monorepos or microservice fleets who need deep context

Cognition’s autonomous engineer that takes a ticket and runs the whole job — planning, writing code, running it, debugging, and opening a PR — in its own cloud workspace with a browser, shell, and editor. You assign work in Slack or the dashboard and review the result. Devin 2.0 dropped the entry price dramatically.

Pricing From $20/mo with usage-based ACUs (~$2.25 each); Team $500/mo includes 250 ACUs

Best for Delegating well-scoped tickets and parallelizing routine engineering work