# AgentVerdict

> Which AI agent should you actually use? Short, sourced, honest comparisons. Independent directory of 34 AI agents and AI coding tools across 6 categories. Every price is sourced from the vendor's page with a "last verified" date (latest: 2026-11-02).

Cite as: "AgentVerdict (https://agentverdict.dev)". Prices change often; check the linked source.

## Categories
- [Coding agents & AI IDEs](https://agentverdict.dev/category/coding-agents/): Agents and editors that read, write, test and ship code in your repo.
- [Vibe-coding app builders](https://agentverdict.dev/category/app-builders/): Describe an app in plain English and get a running, hosted web app.
- [General & computer-use agents](https://agentverdict.dev/category/general-agents/): Assistants that browse, use apps and finish multi-step tasks for you.
- [Research agents](https://agentverdict.dev/category/research-agents/): Agents that search, read and cite sources to answer hard questions.
- [Agent frameworks & automation](https://agentverdict.dev/category/frameworks/): Libraries and platforms for building and running your own agents.
- [Gaming & creative AI](https://agentverdict.dev/category/gaming-creative/): AI for making games, game art, NPC voices and music.

## Tools
- [Claude Code](https://agentverdict.dev/tools/claude-code/): Hard, multi-step coding work in the terminal when you want the highest benchmark scores. Free: No. Claude Code is not on the Free plan. (verified 2026-11-02)
- [OpenAI Codex](https://agentverdict.dev/tools/openai-codex/): Getting strong agentic coding for the lowest cost per task, starting free with a ChatGPT account. Free: Yes, limited Codex access on ChatGPT Free. (verified 2026-11-02)
- [Cursor](https://agentverdict.dev/tools/cursor/): An AI-first code editor for everyday work, with local and cloud agents. Free: Yes, limited agent and chat requests. (verified 2026-11-02)
- [GitHub Copilot](https://agentverdict.dev/tools/github-copilot/): Teams already on GitHub who want agent mode, cloud agents and code review in their existing IDE. Free: Yes. Copilot Free (2,000 completions/mo, auto model selection). Copilot Student is free for verified students. (verified 2026-11-02)
- [Devin (Desktop & Cloud)](https://agentverdict.dev/tools/devin/): Managing several local and cloud coding agents from one IDE, including third-party agents. Free: Yes. Light agent quota, unlimited inline edits and Tab completions. (verified 2026-11-02)
- [Google Antigravity](https://agentverdict.dev/tools/google-antigravity/): Developers on Google's stack who want a free agent-first IDE and CLI powered by Gemini. Free: Yes, $0/month individual tier. (verified 2026-11-02)
- [Kiro](https://agentverdict.dev/tools/kiro/): Structured, spec-first agentic development, especially for AWS-centric teams. Free: Yes. Kiro Free includes 50 credits with access to open-weight models and Claude Sonnet 4.5, with limits. (verified 2026-11-02)
- [Augment Code](https://agentverdict.dev/tools/augment-code/): Small teams working in large codebases who want one flat team price. Free: Unknown. No free plan was stated on the pricing page we read. (verified 2026-11-02)
- [JetBrains Junie](https://agentverdict.dev/tools/jetbrains-junie/): Developers who live in IntelliJ, PyCharm, WebStorm and other JetBrains IDEs. Free: Yes. Junie Lite: a free model, always on, no trial. (verified 2026-11-02)
- [Zed](https://agentverdict.dev/tools/zed/): A very fast editor with built-in AI, where you pay only for the tokens you use. Free: Yes. The editor is free forever. A 14-day Pro trial includes $5 of usage. (verified 2026-11-02)
- [Amp](https://agentverdict.dev/tools/amp/): Agents that keep working on remote machines ("orbs") after you close your laptop, using models you already pay for. Free: Yes. Hobby: unlimited self-hosted orbs, BYOK or your existing ChatGPT/Claude subscriptions, no Amp token fees. (verified 2026-11-02)
- [Cline](https://agentverdict.dev/tools/cline/): An open-source agent in your editor where you control the model and pay only for inference. Free: Yes. The extension is free for individual developers. (verified 2026-11-02)
- [Aider](https://agentverdict.dev/tools/aider/): Terminal pair programming with automatic git commits and any LLM, including local models. Free: Yes. Free and open source; you pay for whichever model you connect. (verified 2026-11-02)
- [OpenCode](https://agentverdict.dev/tools/opencode/): Trying a capable open-source terminal agent for $0, with free models and no account. Free: Yes. "100% free to use", including free models usable without an account. (verified 2026-11-02)
- [Goose](https://agentverdict.dev/tools/goose/): An extensible open-source agent that can install, execute, edit and test with any LLM. Free: Yes. Apache-2.0 open source. (verified 2026-11-02)
- [Lovable](https://agentverdict.dev/tools/lovable/): Non-coders who want to chat their way to a working, hosted web app. Free: Yes. 5 daily build credits (up to 30/mo) plus 20 monthly Cloud credits. (verified 2026-11-02)
- [Bolt](https://agentverdict.dev/tools/bolt/): Quickly prototyping and hosting web apps in the browser, with token-based pricing. Free: Yes. 1M tokens/month (300K daily limit), public and private projects, hosting, Bolt branding. (verified 2026-11-02)
- [v0](https://agentverdict.dev/tools/v0/): Generating polished React/Next.js UIs and apps that deploy straight to Vercel. Free: Yes. 7 messages/day, deploy to Vercel, Design Mode, GitHub sync. (verified 2026-11-02)
- [Replit](https://agentverdict.dev/tools/replit/): Building and deploying full-stack apps with databases from one browser workspace. Free: Free Mode usage is referenced on the pricing page. Free plan details weren't readable on the page we fetched. (verified 2026-11-02)
- [Base44](https://agentverdict.dev/tools/base44/): Quickly building internal tools and simple apps with built-in integrations. Free: Yes. All core features on the Free plan. (verified 2026-11-02)
- [ChatGPT (Work & agent features)](https://agentverdict.dev/tools/chatgpt/): A general assistant that can browse, run deep research, schedule tasks and work in your apps. Free: Yes. Unlimited text chats with GPT-6 (abuse guardrails apply) and limited deep research, Codex and ChatGPT Work on desktop. (verified 2026-11-02)
- [Claude (apps & Claude in Chrome)](https://agentverdict.dev/tools/claude/): Delegating and scheduling multi-step work across your apps, docs and browser. Free: Yes. Chat, web search, files and code, memory, connectors, Artifacts, limited usage. (verified 2026-11-02)
- [Manus](https://agentverdict.dev/tools/manus/): Handing off whole tasks (research, slides, sites, media) to an autonomous agent. Free: Yes. 300 daily refresh credits, 1 concurrent task, 2 scheduled automations. (verified 2026-11-02)
- [Perplexity](https://agentverdict.dev/tools/perplexity/): Cited answers and deep research across the web, files and apps. Free: Yes. Search with cited answers and limited weekly usage. (verified 2026-11-02)
- [Elicit](https://agentverdict.dev/tools/elicit/): Searching and summarizing academic papers and pulling data out of them. Free: Yes. (verified 2026-11-02)
- [LangChain / LangGraph + LangSmith](https://agentverdict.dev/tools/langchain/): Building custom, stateful agents in code and tracing, evaluating and deploying them. Free: Yes. Developer plan $0 (1 seat, 5k base traces/month, 1 free small serverless deployment). (verified 2026-11-02)
- [CrewAI](https://agentverdict.dev/tools/crewai/): Orchestrating teams ("crews") of role-based agents in Python. Free: Yes. The open-source framework is free; the platform's Free tier includes 50 workflow executions/month. (verified 2026-11-02)
- [OpenAI Agents SDK](https://agentverdict.dev/tools/openai-agents-sdk/): A lightweight, code-first way to build multi-agent workflows. Free: Yes. MIT-licensed SDK. (verified 2026-11-02)
- [n8n](https://agentverdict.dev/tools/n8n/): Visual workflow automation with AI agents, priced per workflow run, not per step. Free: Free trial of n8n Cloud (14 days on Business, no card required). (verified 2026-11-02)
- [Rosebud AI](https://agentverdict.dev/tools/rosebud/): Vibe-coding playable browser games by chatting with an AI. Free: Yes. 8,000 credits, refilling weekly in eligible regions. (verified 2026-11-02)
- [Ludo.ai](https://agentverdict.dev/tools/ludo-ai/): Game ideation, market research and asset generation for indie studios. Free: Free trial. (verified 2026-11-02)
- [Scenario](https://agentverdict.dev/tools/scenario/): Generating consistent game art and assets across many models. Free: Yes. 50 free daily credits, no credit card. (verified 2026-11-02)
- [Inworld AI](https://agentverdict.dev/tools/inworld/): Voices and real-time AI characters (TTS, STT, LLM) for games and apps. Free: Yes, start free. (verified 2026-11-02)
- [Suno](https://agentverdict.dev/tools/suno/): Making soundtracks and songs for games, streams and videos. Free: Yes. Free models, no monthly song downloads. (verified 2026-11-02)

## Comparisons
- [Claude Code vs OpenAI Codex](https://agentverdict.dev/compare/claude-code-vs-openai-codex/)
- [Cursor vs GitHub Copilot](https://agentverdict.dev/compare/cursor-vs-github-copilot/)
- [Cursor vs Devin (Desktop & Cloud)](https://agentverdict.dev/compare/cursor-vs-devin/)
- [Lovable vs Bolt](https://agentverdict.dev/compare/lovable-vs-bolt/)
- [Lovable vs v0](https://agentverdict.dev/compare/lovable-vs-v0/)
- [ChatGPT (Work & agent features) vs Manus](https://agentverdict.dev/compare/chatgpt-vs-manus/)
- [Perplexity vs ChatGPT (Work & agent features)](https://agentverdict.dev/compare/perplexity-vs-chatgpt/)

## Best for
- [Best free AI coding agent](https://agentverdict.dev/best/free-ai-coding-agent/)
- [Best tool to vibe-code an app without coding](https://agentverdict.dev/best/vibe-coding-app-no-code/)
- [Best open-source coding agent](https://agentverdict.dev/best/open-source-coding-agent/)
- [Best AI agents and tools for indie game dev](https://agentverdict.dev/best/indie-game-dev/)
- [Best AI research agent](https://agentverdict.dev/best/research-agent/)
- [Best framework to build your own AI agent](https://agentverdict.dev/best/build-your-own-agent/)

## Optional
- [Full text for LLMs](https://agentverdict.dev/llms-full.txt)
- [Structured data (JSON)](https://agentverdict.dev/api/tools.json)
- [Methodology](https://agentverdict.dev/about/)