Review Expert MCP server for AI agents (early access)

No. The MCP server is in early access and not yet generally available. There is no public endpoint or public API today. Email hello@review.expert with the subject "MCP early access" to join the waitlist.

The plan is read access to the App Store and Google Play reviews your team has connected: list apps, fetch and search reviews, get stats such as average rating, rating distribution and response rate, and see rating-based sentiment over time. A reply-draft tool is planned for a later phase. Drafts will never be posted without a person approving them in Review Expert.

Any client that supports remote MCP servers over Streamable HTTP, including Claude (Desktop, web and Claude Code), ChatGPT connectors in developer mode, Cursor and VS Code, and custom agents built with an MCP SDK.

The plan is OAuth 2.1 with Dynamic Client Registration (RFC 7591). The client registers itself, the person verifies their email with a one-time code, Review Expert creates a team on a free agent tier, and the person approves the requested scopes on a consent screen. No sales call is needed.

Yes, that is planned. Team owners and admins would create team-scoped API keys (prefixed rx_live_) under Admin → Settings → API keys and send them as an Authorization: Bearer header. Keys would be stored hashed.

Proposed pricing (subject to change): a free agent tier with 500 tool calls a month, 5,000 calls on Starter, 25,000 on Pro and 100,000 on Enterprise, with overage at $5 per 10,000 calls on paid plans. Today's Review Expert plans are $49, $149 and $399 a month, or $39, $119 and $319 billed annually.

No. All planned tools are read-only except draft_reply, which only returns text. Posting a reply will stay a human action inside Review Expert.

No. Sentiment in Review Expert is derived from the star rating: 4–5 stars is positive, 3 is neutral, 1–2 is negative. Text-based sentiment and topic analysis are planned for later.

Yes. Appbot, AppFollow, AppTweak and Appfigures already offer MCP servers, usually as part of a higher-tier plan or a separate API subscription. Review Expert's plan differs in three ways: agents can sign up themselves, usage is metered per tool call at published prices, and the same team can route reviews to Slack, Linear, Microsoft Teams, Zendesk and webhooks.

Review text is written by the public, so the server will mark it as untrusted data, not instructions, and cap response sizes. Each request is scoped to one team, and the server will only return reviewer details that the store itself exposes.