ASO-Copilot
MCP server for managing App Store Optimization: add apps, track keywords, pull health reports, spy on competitors, find keyword opportunities, and write AI copy suggestions.
README
ASO Copilot
A self-hosted, free alternative to tools like rankd.dev/AppTweak/ASOMobile for developers managing their own apps. Single-user, no accounts, no billing — just your apps.
Also exposes an MCP server so an agent (Claude Code, Claude Desktop, etc.) can manage your ASO directly: add apps, track keywords, pull health reports, spy on competitors, find keyword opportunities, and write AI copy suggestions itself with no separate API key.
https://github.com/user-attachments/assets/970ce32e-a980-42ce-a0eb-529c425439af
Features
- Daily keyword rank tracking — real ranks from live App Store / Google Play search results
- Unlimited apps, keywords, competitors — it's your database, no plan limits
- Competitor spy — track competitor apps and see their rank for every keyword you track
- Health score analysis — title/subtitle/screenshots/description/ratings/freshness breakdown with suggestions, same shape as the paid tools
- Keyword research — "find winning keywords" derives candidates from your + competitors' metadata and scores them by estimated demand/difficulty using live store search data
- Global Reach — scans your rank for a keyword across ~40 major App Store / Google Play storefronts, plotted on a world map
- Review analysis — rating distribution, positive/negative theme extraction, and a keyword-gap finder that mines your own reviews for untracked search terms
- Product health overlay — pairs your app's real daily active users (via a per-app PostHog project) against its ASO health score over the same window
- AI Copy Suggestions — rewrites title/subtitle/description for 8 locales (US/UK/Canada
English, Spanish, German, French, Portuguese, Japanese), adapting keywords into real local
search terms rather than translating literally. Two ways to generate it, no lock-in to either:
- via OpenRouter (any backing model) through the web UI, or
- key-free, by pointing any MCP-connected model (e.g. Claude Code) at the
prepare_copy_localization_brief/save_copy_suggestionstools — the model composes the copy itself, no API spend beyond what you're already paying for the agent
- MCP server at
/api/mcpfor agent-driven ASO management
Stack
Next.js (App Router) + Prisma + PostgreSQL. Store data comes from the public iTunes Search/Lookup
API (App Store) and google-play-scraper (Google Play). There is no free official API for search
volume or keyword rank position, so:
- Rank position is real: it's your app's actual index in a live search for that term.
- Volume/difficulty are heuristics (0-100) derived from result-count and competitor-authority proxies, not real search volume — there's no free source for that.
- iOS subtitle is scraped from the public App Store product page (not exposed by the iTunes
API). If Apple changes their page markup this can silently start returning
null— seefetchSubtitleinsrc/lib/stores/appstore.ts. - Every store call (both platforms) retries transient failures (429s, 5xx, network errors) with
backoff via
src/lib/withRetry.ts, so one rate-limited request doesn't fail an entire daily tracking pass.
Setup
npm install
docker compose up -d # local Postgres on port 5433
npx prisma migrate dev # create tables
npm run dev
Open http://localhost:3000, click Add App, search for your app, and add it.
Copy .env.example to .env and adjust DATABASE_URL if you're pointing at a different Postgres
instance (e.g. Supabase/Neon/RDS instead of the bundled docker-compose one).
Daily tracking
Rank checks don't run automatically — trigger POST /api/track (protected by CRON_SECRET from
your .env) once a day:
curl -X POST http://localhost:3000/api/track -H "Authorization: Bearer $CRON_SECRET"
scripts/daily-track.sh wraps this for cron:
0 6 * * * APP_URL=http://localhost:3000 CRON_SECRET=... /path/to/scripts/daily-track.sh >> /var/log/aso-track.log 2>&1
Any external scheduler works the same way (GitHub Actions schedule: trigger, a hosting
provider's cron jobs, etc.) — it just needs to hit that one endpoint once a day.
Or run the bundled cron container
docker compose up -d cron starts a small Alpine container (docker/cron/) that runs
scripts/daily-track.sh on a schedule via crond, hitting the app over host.docker.internal
since the app itself runs on the host, not in Docker. Configure it via .env:
CRON_SECRET="..." # required, shared with the app's own .env
APP_URL="http://host.docker.internal:3000" # optional, this is the default
CRON_SCHEDULE="0 6 * * *" # optional, standard 5-field cron syntax
Connecting an agent (MCP)
The app exposes an MCP server at /api/mcp (Streamable HTTP transport) with tools for listing
apps, searching stores, adding apps/keywords/competitors, pulling health reports, finding winning
keywords, generating AI copy suggestions, and running a tracking pass on demand.
Point any MCP-compatible client at it, e.g. in Claude Code (.mcp.json, already checked in at the
repo root pointing at localhost):
{
"mcpServers": {
"aso-copilot": {
"url": "http://localhost:3000/api/mcp"
}
}
}
There's no auth on this endpoint by default — it has full read/write access to your ASO data. Set
MCP_AUTH_TOKEN in .env before exposing it beyond localhost, then have your client send it:
{
"mcpServers": {
"aso-copilot": {
"url": "http://localhost:3000/api/mcp",
"headers": { "Authorization": "Bearer <MCP_AUTH_TOKEN>" }
}
}
}
AI Copy Suggestions
Two independent ways to generate rewritten title/subtitle/description copy, in
src/components/AICopySuggestions.tsx on each app's Overview tab - pick either, both write to the
same AiCopySuggestion table so the panel shows whichever ran most recently:
- OpenRouter (works from the web UI): set
OPENROUTER_API_KEYin.env(get one at openrouter.ai/keys); optionallyOPENROUTER_MODELto pick a different model than the default. Click Get AI suggestions. - MCP, no key needed: have an MCP-connected model call
prepare_copy_localization_brief(returns current copy, character limits, and a localization instruction per locale - no network call, no key) and write the copy itself, thensave_copy_suggestionsto persist it. Useful if you're already paying for an agent and don't want a second LLM bill for this.
Analytics (optional)
Two separate PostHog integrations - don't mix them up:
<img width="1558" height="521" alt="image" src="https://github.com/user-attachments/assets/04985484-c04c-4e2d-95c4-65c56c2f9e1c" />
- This tool's own usage analytics (are you clicking around this app, not any tracked app's
users) — off by default. Set
NEXT_PUBLIC_POSTHOG_KEY(and optionallyNEXT_PUBLIC_POSTHOG_HOSTif self-hosting PostHog) in.envto enable pageview tracking viasrc/components/PostHogProvider.tsx— no code changes needed, it's a no-op until the key is set. - Per-app product health — each tracked app can link its own PostHog project (Settings tab,
src/components/PostHogSettings.tsx) so its real daily active users show up next to its ASO health score on the Overview tab. Unrelated to the key above; configured per-app in the UI, not via env vars.
Project layout
src/lib/stores/— App Store (iTunes) and Google Play data fetchingsrc/lib/health.ts— health score enginesrc/lib/research.ts— keyword candidate generation + volume/difficulty scoringsrc/lib/ai.ts/src/lib/aiLocales.ts— AI Copy Suggestions: OpenRouter prompt + call, the key-free localization-brief builder, and the 8-locale listsrc/lib/reviewAnalysis.ts— rating distribution + positive/negative theme extractionsrc/lib/posthogIntegration.ts— per-app PostHog client (daily active users query)src/lib/storeLinks.ts/src/lib/metricColor.ts— small shared UI helpers (outbound store links, volume/difficulty coloring) kept out ofresearch.tsso Client Components importing them don't drag in server-only store-scraping dependenciessrc/lib/appService.ts— shared logic used by both the REST API and the MCP toolssrc/app/api/— REST routessrc/app/api/[transport]/route.ts— MCP serverdocker/cron/— optional bundled cron container for daily tracking (see "Daily tracking" above)
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.