# Appllama — MCP server card for AI agents

Study screens, flows and paywalls from top-earning iOS apps, then build from what wins.

Registry name: `io.appllama/appllama` · category: Other · version 1.0.1 · updated 2026-08-27

## Connect
- Remote (Claude Code): `claude mcp add --transport http appllama https://mcp.appllama.io/mcp`

## Facts
- Repository: https://github.com/Appllama/appllama-skills (5,227 stars)
- Website: https://appllama.io/mcp

Before you connect a server: read its repository, and give it only the permissions and keys it needs.

## Links
- AgentGid page: https://agentgid.com/mcp/servers/io-appllama-appllama/


## Appllama: study top-earning iPhone app screens before building

Appllama is a paid library of real screens from top-earning iPhone apps, with their revenue and download numbers. Its connector lets your AI assistant search those apps, walk through their screens and study onboarding and paywalls, then build your app from what it found. It needs an Appllama Pro plan.

### Where to use it
- Study the leaders before building: Before your assistant builds a habit tracker, it looks up the top-earning habit apps and walks their screens one by one.
- Plan onboarding and the paywall: See how many steps winning apps put in onboarding, what each step asks and where the paywall appears.
- Improve one screen you already have: Paste a screen id copied on appllama.io or a screenshot, and the assistant pulls similar screens from strong apps to compare.
- Learn how leaders build one element: Look at how top apps draw a tab bar, a progress bar or an empty state before your assistant builds its own.
- Start from your saved boards: If you already saved screens on appllama.io, the assistant opens that board first instead of searching again.

### How to start
- Get an Appllama Pro plan; the connector is part of it.
- Add the address as a custom connector in your assistant and approve it with your Appllama account.

```
https://mcp.appllama.io/mcp
```

- Start each session with get_credits, which costs nothing, to see what is left.

### License
The connector is paid (Pro) and licensed to one person for research, reference and design inspiration. The screens belong to the apps' owners, so do not copy them, republish them or train models on them.

### Good to know
- Every tool call spends one credit from 1,500 a month; only the credit check is free. Credits reset on the 1st and do not roll over.
- Image links expire after about an hour and are meant for looking at, not for keeping.
- Sweeping the whole library is against the terms and is detected; accounts that do it are restricted.
- Every image carries a small Appllama watermark in the top-left corner; never copy it into your own design.
- A Pro plan is for one named person; sharing the login is not allowed, teams need a team account.

Checked 2026-11-10 against: https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/README.md, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/.mcp.json, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/mcp.json, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/.claude-plugin/plugin.json, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/.cursor-plugin/plugin.json, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/skills/appllama-usage/SKILL.md, https://github.com/Appllama/appllama-skills/blob/dd5caaec3d5d50ad7fc0324da238119c6b7c3707/LICENSE, https://appllama.io/mcp, https://appllama.io/llms.txt, https://appllama.io/terms, https://appllama.io/pricing

### Mistakes to avoid
- Don't: Downloading every studied screen into a local research/img folder, as the vendor's own usage skill suggests → Do: Keep notes and screen ids; fetch a fresh link from the id when you need to look again (The terms say media links are for immediate study, not storage, and forbid caching or archiving the library.)
- Don't: Looping over every page of search results to collect the whole catalog → Do: Study the few apps that matter for the task (Bulk retrieval is detected and gets the account restricted, and every call spends a credit.)
- Don't: Jumping to page 3 of a list or reusing a cursor with a new query → Do: Page through in order, passing next_cursor back with the same query; restart from page one if a cursor errors (Pagination only goes forward, one page at a time.)
- Don't: Rebuilding a screen with the Appllama watermark or another app's logo in it → Do: Rebuild the pattern in your own design (The watermark is provenance and the logos belong to the apps' owners.)

### Questions
- Is it free? No. The connector is part of the paid Pro plan, with 1,500 credits a month; browsing the site has a free part.
- Can I copy the screens into my app? No. You may study them for research and inspiration and build your own; the screens and logos stay with the apps' owners.
- Can I use the screens to train an AI model? No. The terms forbid training, fine-tuning, evaluating or indexing a model with the library.
- Do I have to install the skills? No. The connector works on its own; the free skills only teach an assistant how to research with it and how to build native screens.

---
AgentGid's catalog is curated: every entry was checked against the maker's official pages (registry, repository, license and pricing pages) on the date shown; a license value shows the quote it stands on wherever AgentGid has one. Usage numbers come from AgentGid's own scan of public repositories set up for AI coding agents. Claims by makers are reported, not tested by AgentGid.
