AsoBeast is an app-store optimization toolkit for developers who want to monitor how their apps appear in Apple App Store and Google Play searches. It collects daily keyword positions, competitor listings and reviews into a dashboard you can host yourself. Its 1.9.0 release, published October 10, adds useful workflow changes rather than just a new product name.
- What’s new: Separate listing snapshots for each tracked market; CSV keyword import with a preview; and local text recognition for App Store screenshot captions. Caption text is treated as a weak signal, not proof that Apple indexes a keyword. Google Play screenshots are tracked for changes but their text is not read.
- Agent interface: Model Context Protocol (MCP), a standard interface for connecting assistants to tools, exposes the stored data through HTTP or a local stdio process. Personal tokens are read-only by default. Explicit write scope unlocks five operations: track/untrack keywords, add/remove competitors, and change an action’s status. That is not permission to rewrite your store listing or manage billing.
- Setup: The supported Docker Compose stack runs a Next.js dashboard, NestJS API, PostgreSQL and Redis. Generate the database password and authentication secret, start the stack, then create the owner account at
http://localhost:3001. Published container images are available if you do not want to build from source. - Cost: The self-hosted edition is free software under AGPL-3.0, with no required paid ASO data-provider key. Hosting, maintenance and any optional model API or proxy service are separate costs. A hosted service also exists; its subscription price is not verified here.
The practical limit is collection, not dashboard polish. Public store endpoints can change or throttle your IP, and popularity/difficulty scores are estimates—not measured acquisition outcomes. More markets mean more requests; the project’s capacity documentation explicitly says some larger mixed-store workloads cannot finish within a daily window on one address.
For upgrades, use the 1.9.0 Compose file, not just a new image tag: screenshot text recognition raises the API memory limit to 1 GB. Back up PostgreSQL first; release migrations also clean stored metadata and untrack some incorrectly extracted category keywords while retaining their history.
This is a documentation-based capture, not a hands-on validation of ranking accuracy or installation reliability. Start with a small keyword set and a read-only agent token.