The right AI tool for the job, ranked.
Use-case rankings for vibe coders and builders. Every list is scored on hands-on tests, momentum, value and freshness, and dated so you know it's current.
Categories
Pick the job. We rank the tools. Prefer one page? See all rankings.
App builders
Prompt-to-app platforms that ship a working full-stack app from a description.
AI code editors
IDEs with agents, autocomplete and codebase-aware chat.
CLI coding agents
Terminal agents that read, edit and run your repo.
Agent frameworks & MCP
Building blocks for your own agents and tool integrations.
Code review & testing
Catch what the agent got wrong before it ships.
Deploy & hosting
Ship it: hosting, databases and edge platforms.
New this week
Verified launches and updates from the last 7 days. Listed, not yet ranked. Every item has a source link.
- Mintlify Autopilot Agent frameworksAlways-on docs agent: watches your code, drafts doc updates when something changes and opens a pull request for review. Product Hunt launch. Source
- Jotform App Builder for ChatGPT & Claude App buildersBuild and edit no-code apps (portals, hubs, catalogs) inside ChatGPT or Claude, then share them without leaving the chat. Product Hunt launch. Source
- Composio Instant Agent frameworksLets AI agents pay for their own paid tools per call through MCP. Product Hunt launch. Source
- MiniMax M3.1 Flash on Vercel AI Gateway ModelsOpen-weight multimodal model for coding and agent work, 1M context; free on AI Gateway until Oct 22. Source
- Mercury 3 on Vercel AI Gateway ModelsInception's diffusion text model, 1,000+ tokens per second and 1M context, at 50% off launch pricing. Source
- Superset Plugins AI code editorsInstall an MCP, skill or login once and every agent (Claude Code, Codex, Cursor) gets it on every machine. Source
- Beam.win CLI coding agentsRun AI agents on one Mac and use its Terminal and browser windows natively on another, encrypted. Product Hunt launch. Source
- Framefields Agent frameworksOpen-source, code-first video engine for coding agents: motion graphics, 3D and audio rendered to MP4. Product Hunt launch. Source
- Sphinx Code reviewFree open-source hook that quizzes you on your AI agent's code before you push to production. Show HN. Source
- WhatsApp MCP on Cloudflare Agent frameworksUnofficial read-only MCP server so your agent can search your own WhatsApp messages; runs on Cloudflare Workers. Show HN. Source
- shim-cli CLI coding agentsOpen-source CLI that shows what Claude Code, Codex and Copilot CLI send, and masks secrets where it can. Product Hunt launch. Source
- DeepSeek V4.1 Flash 1017 ModelsRefreshed DeepSeek Flash model, now also on Vercel AI Gateway with image input. Source
- Perplexity SPACE Deploy & hostingPerplexity open-sourced its sandbox platform for running agent code in isolated Firecracker microVMs. Source
- Ultracite v8 Code reviewZero-config linter and formatter for agent-written code; writes shared rules into AGENTS.md for 40 agents. Product Hunt launch. Source
- Craft Agents AI code editorsOpen-source local desktop for agents: bring your own model, plus MCP, APIs and a built-in browser. Product Hunt launch. Source
- NitroStack Agent frameworksOpen-source TypeScript SDK and platform to build, test and deploy MCP servers. Product Hunt launch. Source
- Cursor cloud agent Canvases AI code editorsCursor cloud agents can build interactive canvases and share them by link; Environments API in beta, up to 300 repos per environment. Source
- Harness by Autonomous CLI coding agentsOpen-source command center for Claude Code, Codex and other coding agents across machines: desktop app, terminal and mobile, end-to-end encrypted. Product Hunt launch. Source
- Parallax AI code editorsOne app to run Claude Code, Codex and 40+ coding agents on your own subscriptions, with one-click PR merge and auto fallback when a plan runs out. Product Hunt launch. Source
- Pomerado Agent frameworksTurns websites without APIs into hosted MCP servers or APIs your agents can call; open source. Product Hunt launch. Source
- Muse Spark Open (Meta) ModelsMeta's open-weight coding and agent model: 816B total / 42B active, 1M context. Weights on Hugging Face; APIs on OpenRouter, Groq, Together and more. Source
- Claude Code Review (GA) Code reviewAnthropic's multi-agent PR reviewer is out of research preview for Team and Enterprise; most reviews in 6-10 minutes, Quick/Standard/Deep per repo. Source
- GitHub agentic autofix (GA) Code reviewAgentic autofix for code scanning alerts is generally available: explores the codebase and proposes fixes for security alerts. Source
- Vercel Daemons (public beta) Deploy & hostingRun long-running queue consumers, workflow workers and self-hosted CI runners on Vercel without an HTTP server. Source
- Open Scion Agent frameworksOpen-source version control built for coding agents; 1,000+ GitHub stars in its first day. Source
- withHuman Agent frameworksOpen-source approval layer that intercepts agent tool calls (Claude Code, Codex, LangGraph, Pydantic AI) and routes risky ones to a human. Product Hunt launch. Source
- Navige Agent frameworksCuts AI agent costs while jobs run: routes to smaller models, caches reads, sets budgets and catches runaway loops. Product Hunt launch. Source
- Polylane Impact Analysis Code reviewFlags risky pull requests before they reach production by linking code, infra and observability data. Product Hunt launch. Source
- Agent Vault (Self) Agent frameworksLets agents like Claude Code, Codex and Grok Bot log in and pay by card without seeing your real card or passwords; spending caps and approvals. Source
- RDST (Readyset) Code reviewOpen-source local agent that watches Postgres/MySQL for query performance regressions. Show HN. Source
- Buddy Sandboxes Deploy & hostingIsolated persistent Linux VMs for AI coding agents, with Docker, SSH, public URLs, MCP/CLI/SDK. Launched on Product Hunt. Source
- CodeCrab Code reviewLocal-first AI code review desktop app; integrates GitHub, GitLab, Claude, Codex, Cursor and Copilot CLI. Product Hunt launch today. Source
- Nemotron 3.5 Lightning on Vercel AI Gateway ModelsNVIDIA's open-weight 30B MoE (3B active) free on Vercel AI Gateway as nvidia/nemotron-3.5-lightning-free, 1M context. Source
- Graphicle Agent frameworksOpen-source Rust code graph and memory for coding agents (Claude Code, Codex, Cursor, any MCP client). Show HN. Source
- Grok Voice 3.0 ModelsxAI speech-to-speech model in the Voice Agent API, $0.08/min (same price as 2.0). Source
- Lovable growth tools App buildersLovable now plans and runs growth campaigns (Google, YouTube, Meta) for your app. Source
- Amazon S3 in Vercel Marketplace HostingS3 with guided setup and auto project config in Vercel and v0. Source
- Vani Code Agent for VS Code AI code editorsFree VS Code extension: coding agent with diff review and approved command runs. Source
- Base Code App buildersBrowser interface for non-engineers to request and review changes to an existing codebase. Source
- Gemini 4 Argon ModelsGoogle's new frontier model, 1M unified token window, intro $2/$10 per M tokens until Nov 30. Source
- Cursor Origin App Marketplace AI code editors20+ deploy, CI, review and security apps for Origin; open to public submissions. Source
- Gemini 4 Argon in GitHub Copilot AI code editorsGoogle's new frontier model now GA in Copilot (Pro+, Max, Business, Enterprise). Source
Rankings never for sale.
Paid placement never changes a rank. Sponsored items are always labeled. Every ranking shows its last-reviewed date. The scores you see today are preview scores, not yet tested. They are editorial estimates until our hands-on test suite has run.