RankedFor
AI tool rankings, without the catalog

We test the tools. We pick one for each job.

New releases only make the list if they beat what's already there. Rankings for people into AI agents, vibe coding, tech and games, with a clear pick, who should skip it, and what it really costs.

  • One pick per job
  • Rankings are never sold
  • Dated re-tests
  • Free tiers called out
See the rankingsHow we test

Rankings by job

Last updated: 14 Nov 2026
Vibe coding

Ship a vibe-coded app online

Get the app you built with AI live with a domain, a database and payments.

Our pick Cloudflare Workers & Pages · 5 tools compared
Vibe coding

AI coding agent for a real repo

An agent that can read an existing codebase, make changes and open a pull request.

Our pick Claude Code · 4 tools compared
Vibe coding

AI app builder for non-coders

Describe the app in plain English and get something you can click, share and publish.

Our pick Lovable · 5 tools compared
AI agents

Build a customer-support, WhatsApp or email agent

An AI agent that answers customers on chat, email or WhatsApp and hands off to a human.

Our pick Intercom Fin · 4 tools compared
Gaming

AI image and asset generation for indie games

Sprites, icons, textures, concept art and 3D props that match your game's style.

Our pick Scenario · 4 tools compared
Gaming

AI tools for gamers (guides, builds, patch notes)

Get better builds, matchup advice and setup help without tabbing out for an hour.

Our pick Mobalytics · 3 tools compared
Productivity

Meeting notes and transcription

Accurate transcripts, useful summaries and action items, without babysitting a bot.

Our pick Granola · 4 tools compared
Productivity

AI research and search assistant

Ask a question, get a sourced answer or a multi-page report you can actually check.

Our pick Perplexity · 4 tools compared

New this week

All new releases →

Fresh AI tool launches we're watching. They only join a ranking if they beat the current pick.

  • Claude Meetings

    Adds Claude to your call so it can take notes and start work while you talk.

    Via Product Hunt · Meetings, Agents · Up against: Meeting notes and transcription
    Not yet tested

  • Gemini 4 Hydrogen

    Google's fast Gemini model aimed at agentic development for coders.

    Via Product Hunt · Models, Coding · Up against: AI coding agent for a real repo
    Not yet tested

  • Crossfade

    Lets your coding agent design directly in your open Figma file.

    Via Product Hunt · Vibe coding, Design
    Not yet tested

  • Yogurt

    Desktop app (Mac, Windows, Linux) that records calls without a meeting bot; transcripts and summaries run locally via whisper.cpp and llama.cpp. MIT licensed, no account required. Speaker names from Zoom, Teams and Meet, and calendar access, are Mac-only for now. Setup downloads 5–14 GB of local models.

    Via Hacker News (Show HN) · Meetings, Open source · Up against: Meeting notes and transcription

    14 Nov 2026: We read the public README and GitHub metadata (MIT, 1,250 stars). The makers' claims match the README: no bot joins the call, audio/transcripts/summaries stay on the device after a one-time model download, and no account is required. The README does not itself use the phrase 'Granola alternative' (that was the Show HN framing). We did not install or run Yogurt on our meeting-notes test recordings.

    README inspected

  • Modern Web Guidance (Google Chrome)

    Open-source skill and CLI (npx modern-web-guidance) stewarded by the Chrome, Edge, Firefox and MDN teams. It searches and retrieves token-efficient modern web-platform guides and can scan a repo for legacy patterns.

    Via Hacker News (Show HN) · Vibe coding, Open source · Up against: AI coding agent for a real repo

    14 Nov 2026: We ran the public CLI with no login: `npx modern-web-guidance@latest search "animate a dialog modal backdrop"` returned relevant guides including animate-to-from-top-layer; `retrieve animate-to-from-top-layer` returned a 10.7 KB expert guide using @starting-style, overlay and transition-behavior: allow-discrete; `legacy-scan /workspace/rankedfor` scanned 33 guides and flagged one dark-mode modernization in src/styles/global.css. GitHub metadata: Apache-2.0, 10,767 stars, last push 10 Nov 2026. This verifies what the tool is; it is not a full coding-agent ranking test.

    CLI inspected