The ASO MCP server for the AI you already use
The Applyra MCP server connects App Store and Google Play keyword data to Claude, Cursor, Codex, VS Code and any other MCP client. Ask what to target and the agent reads your live ranks, scores a keyword, compares you to a competitor, or starts tracking a term. Twenty tools, both stores, no dashboard to open.
Claude
Cursor
Codex
VS Codeand any MCP clientWhich keywords could my habit tracker realistically rank for?
What an ASO agent answers
Keyword work does not slip because the data is hard to read. It slips because checking it means stopping what you were doing, opening a tab and remembering what you were looking for. Here is what the same four questions look like when the agent can fetch the answer while you stay where you are.
Am I actually gaining on the keywords I care about?
Where do I stand against the apps I track?
Break my category into niches and tell me which one is open.
34 keywords
21 keywords
27 keywords
52 keywords
Who moved in the US productivity chart this week?
- #3Clearlist4
- #6Taskbird0
- #8Dayplan11
- #14Focusly5
Twenty ASO tools, and what each one is for
The App Store and Google Play are covered by the same tools, so a question never has to be asked twice. Your assistant picks the one it needs from the description; you never type a tool name yourself.
Track and rank
Your apps, your keywords, and how their positions move day by day on both stores.
- list_applications
- Every tracked app with store metadata, rating and keyword count
- add_application
- Start tracking an app from its store bundle ID
- list_keywords
- Tracked keywords with current rank, difficulty and traffic
- track_keywords
- Track up to 20 new keywords in one call
- untrack_keyword
- Stop tracking a keyword
- set_keyword_favorite
- Flag the keywords that matter most
- get_keyword_rank_history
- Daily rank evolution over any date range
Research a keyword
Score a term before you spend a release cycle on it, and find the ones you have not thought of.
- inspect_keyword
- Difficulty, traffic, KEI, the top 20 apps and related terms
- list_keyword_inspections
- Everything you have already scored
- run_autocomplete
- What the stores suggest as people type
- list_autocomplete_history
- Past autocomplete queries
- run_niche_analysis
- Cluster a topic into sub-niches with opportunity scores
- list_niche_analyses
- Past niche analyses
Watch competitors
Who sits above you, on what, and whether the gap is closing or opening.
- list_competitors
- Competitor pairs with side-by-side visibility scores
- add_competitor
- Add a rival app by its store bundle ID
- remove_competitor
- Drop a competitor you no longer track
Market and account
The wider store picture, and where you stand against your own plan.
- top_charts
- Top apps by store, country, category and collection, with daily movement
- list_top_chart_categories
- Categories and collections each store supports
- get_app_score_history
- Your visibility score, day by day
- get_account_usage
- Current usage against your plan limits
Install the MCP server in under a minute
Generate an API key in your dashboard, paste the snippet for your client, restart it. The server runs over stdio through npx, so there is nothing to host, nothing to build and no OAuth dance to sit through.
~/Library/Application Support/Claude/claude_desktop_config.json{
"mcpServers": {
"applyra": {
"command": "npx",
"args": ["-y", "@applyra/mcp-server"],
"env": {
"APPLYRA_API_KEY": "your_api_key"
}
}
}
}The server is MIT licensed and open source, so you can read exactly which endpoints it calls before you install it.
Connected? The agentic ASO playbook has the prompts worth pasting first, and the four places an agent will get ASO wrong.
Prefer to drive it yourself? The same data is available through the REST API on the paid plan, with 10,000 requests a month. And if you are still deciding whether an agent belongs in your ASO workflow, the guide to getting your app recommended by ChatGPT covers the other half of the problem.
Questions people ask before connecting
- What is an ASO MCP server?
- MCP, the Model Context Protocol, is an open standard that lets an AI assistant call external tools. An ASO MCP server exposes App Store Optimization data, keyword ranks, difficulty scores, competitor positions, as tools your assistant can call while you talk to it. Instead of opening a dashboard, you ask a question and the assistant fetches the live answer.
- Which AI clients work with the Applyra MCP server?
- Any MCP-compatible client. Applyra publishes setup snippets for Claude Desktop, Claude Code, Cursor, OpenAI Codex, VS Code with GitHub Copilot, and Windsurf. The server runs over stdio through npx, so a client that speaks MCP can use it without any extra work on our side.
- How many tools does the Applyra MCP server expose?
- Twenty. They cover tracked apps and keywords, rank history, keyword inspection with difficulty and traffic scores, autocomplete mining, niche clustering, competitor pairs, top charts, and account usage. Both the App Store and Google Play are covered by the same tools.
- What can an AI agent actually do with my ASO data?
- Read your live keyword ranks and their history, score any keyword for difficulty and traffic before you commit to it, cluster a niche into sub-niches with opportunity scores, compare your visibility against tracked competitors, and read top charts movement by store, country and category. It can also start or stop tracking a keyword and add a competitor.
- Do I need a paid plan to use the MCP server?
- Yes. The MCP server authenticates with an Applyra API key, and API access is part of the paid plan at $9.99 per month. The free plan lets you track an app and five keywords in the dashboard first, so you can see the data before you connect an agent to it.
- Can the agent change my app store listing?
- No. The Applyra MCP server never touches your store listing. It can add or remove things inside your Applyra account, such as tracking a keyword or adding a competitor, and it can read your data. Publishing metadata to App Store Connect or Google Play stays a deliberate step you take yourself.
- Is the Applyra MCP server open source?
- Yes, it is MIT licensed and published on npm as @applyra/mcp-server, and it is listed in the official Model Context Protocol registry and on the Glama MCP directory. You can read exactly what it sends and receives before you install it.
Put your rankings where you work
Start on the free plan, track your first keywords, then connect the agent when the data is worth asking about.