The First Agentic Framework of the PHP Ecosystem | Free | - **Flexible Architecture**: Build everything from simple RAG applications to advanced multi-agent pipelines using modular components that adapt to your specific project requirements.
- - **Multi-Provider Support**: Connect with leading AI providers such as OpenAI, Anthropic, Gemini, Pinecone, and Elasticsearch. You can switch between providers with minimal changes to your code.
|
Shareable docs your AI writes and your team comments on | Freemium | - **AI-written Docs, Sheets, and Slides**: Your AI writes documents, builds decks, and publishes tables. Readers can comment on any line or slide, and they can search, sort, and comment on any cell in a sheet.
- - **Comment-to-Revision Loop**: Readers leave anchored comments in the browser. Your agent reads them, replies, resolves them, and publishes the next version, and you can hand comments over with "@agent" or "Send to agent".
|
Checkout API that lets your AI agent shop real stores | Free tier | - **One API for Many Stores**: Search, carts, checkout, and order lookup work the same way at every supported merchant. You switch stores by changing the `merchant_id`.
- - **MCP Server and REST API**: Connect through 17 MCP tools at `mcp.getagentmall.com` or call the REST API directly. The quickstart includes setup steps for Claude Code and Cursor.
|
Find and fix failures in your production AI agents | Freemium | - **Live Conversation Scoring**: Sampled production requests are scored pass or fail as they flow through the proxy. You track quality on real traffic, and your app's response time doesn't change.
- - **Failure Modes by Root Cause**: Issues are grouped into named failure modes and ranked by severity and frequency. You see what to fix first instead of scrolling through raw logs.
|
One library for all your AI coding agent skills | Freemium | - **Central skills library**: Import skills from your computer, GitHub, GitLab, or Explore. Skillbox remembers where each skill came from, so you have one place to find it again.
- - **Sync to your tools**: Turning on a skill delivers it to the folders your tools read, such as `~/.claude/skills` and `~/.cursor/skills`. You can apply it everywhere or only inside one project.
|
Connect real-world sensors to your AI agents | Freemium | - **Flexible Device Pairing**: Connect a Raspberry Pi or other Linux machine with the `fizz` CLI, pair a phone through a browser link, or post readings from any HTTPS-capable device. Sample sensor streams let you explore before your hardware arrives.
- - **Fizz Assistant**: Ask questions about your stored readings in plain language, typed or spoken. Fizz can also add sensors, walk you through pairing, and create alerts when you ask.
|
An AI product team that plans, builds, and ships for you | Free tier | - **Six specialized agents**: A Project Manager, Architect, Engineer, Tester, Marketing agent, and Community agent each handle their own part of the product lifecycle. All six roles come with every plan, including Free.
- - **Mission control**: A live activity feed shows what each agent is doing, what it has finished, and what it needs from you. You also chat with the PM here to set goals.
|
AI assistant that answers only from your own content | Free tier | - **Answers with source citations**: Every answer links to the page it came from, so you and your users can check it quickly.
- - **Says "I don't know" instead of guessing**: When your content doesn't cover a question, the assistant says so rather than inventing an answer.
|
A design canvas any AI agent can build on | Free tier | - **Agent-Built Designs**: Ask an AI agent to create social posts, carousels, posters, flyers, websites, or app screens, and watch it build them on the canvas. The agent checks its own work and leaves real layers you can edit.
- - **One MCP Address for Every Agent**: Connect Claude Code, Codex, Cursor, ChatGPT, Gemini CLI, GitHub Copilot, Windsurf, OpenCode, or any MCP-compatible agent through a single address. The built-in terminal comes already connected.
|