⌨️ Vibe coding
Ship apps by describing them. Editors, agents, and prompt-to-app builders.
Cursor
AI code editor with agents that work across your whole repo.
Best when you already have code and want an agent that edits many files, runs commands, and keeps context.
BEST FOR: BUILDERS WHO READ CODEFree plan · paid tiers
Claude Code
Terminal-first coding agent from Anthropic.
Strong at long, multi-step changes from the command line. Pairs well with any editor.
BEST FOR: TERMINAL PEOPLENeeds a paid Claude plan or API
Codex
OpenAI's coding agent for CLI, IDE, and cloud tasks.
Hand off a task, let it run in the background, and review the diff.
BEST FOR: PARALLEL BACKGROUND TASKSIncluded with ChatGPT plans · check site
Lovable
Prompt-to-app builder with hosting and a database baked in.
Fastest path from idea to a working, shareable web app if you never want to see a terminal.
BEST FOR: NON-CODERS SHIPPING MVPSFree plan · paid tiers
Bolt.new
Full-stack apps in the browser from a prompt.
Instant in-browser dev environment. Great for prototypes and demos.
BEST FOR: QUICK PROTOTYPESFree plan · paid tiers
New in vibe coding this week
DeepSeek V4.1 Flash (1017 update): DeepSeek pushed a fresh snapshot of its cheap, fast open model and swapped it into its app, site, and API. Open weights, MIT license, with better coding and agent benchmark scores than the last Flash. source ↗
Reflection AI Beam: A big open-weight model aimed at coding and long agent runs, with a 1M-token context window and day-one support in Ollama and vLLM. It ships alongside Mirror, an open-source agent harness. source ↗
Gemini 4 Argon: Google's new frontier model is now global for developers and subscribers. Big jumps in coding persistence, plus a 1M-token output limit. source ↗
Rankings are editorial and can't be bought. Sponsored placements are always labeled. How we rank →