# Public MCP tool map The complete set of AppFollow MCP tools available to this toolkit, by base name — a host may add its own prefix, so match the names. Tool names are validated for parity in CI. Credit costs live in `references/costs.md`. ## Service | Tool | Purpose | |---|---| | `whoami` | identity / session check | | `get_server_info` | deployed build sha + liveness | | `get_credits` | remaining credit balance | ## Apps & collections | Tool | Purpose | |---|---| | `list_collections` | list the account's collections — `collectionId` is the `apps_id` other tools take, `collectionCode` the `collection_name` that `get_reviews_ai_summary` takes | | `list_apps` | list the apps of one collection (`apps_id` required; dedupe target) — `extId` is the `ext_id` other tools take | | `add_collection` | create a collection (preview→confirm) | | `add_collection_preview` | mint a confirmation token for `add_collection` | | `add_app` | track an app in a collection — `store` (short code), `ext_id`, `apps_id`, `locale` (two-letter country code); optional `is_competitor` (default `false`: one of the user's own apps), `has_translation` (preview→confirm) | | `add_app_preview` | mint a confirmation token for `add_app` | | `remove_app` | remove an app from a collection — `store` (short code), `ext_id`, `apps_id`; the inverse of `add_app`, frees the app or competitor slot; the lookup includes archived apps, so it removes an archived app too; an app not in the collection fails with `NOT_FOUND` before any write and is not charged; show the user the `list_apps` row first, and for an archived app (not listed there) its `ext_id` and `store` (preview→confirm) | | `remove_app_preview` | mint a confirmation token for `remove_app` | `list_apps` reports `store` as the long name, while the `store` argument takes the short code (`report_review_concern` alone takes the long name): `itunes` = `as`, `googleplay` = `gp`, `microsoftstore` = `ms`, `amazon` = `am`, `macstore` = `mc`, `xstore` = `xs`, `samsung` = `sg`, `trustpilot` = `tp`, `huaweiappgallery` = `hg`, `steam` = `st`. `add_app`, `remove_app`, `get_ratings_history` and `reply_to_review` take all ten codes; `get_reviews_ai_summary` takes `as`, `gp`, `ms`, `am`. ## Reviews | Tool | Purpose | |---|---| | `get_reviews` | paged raw reviews of one collection (workspace) — `apps_id`, `from`, `to` required; `ext_id` narrows to one app (absent = the collection's own apps; its competitor apps are left out); optional `country`, `lang`, `q`, `rating`, `sort`, `limit` (1–100 per page); page by passing each `nextCursor` back as `cursor` until it is null or absent (evidence); optional `testing_feedback` for Google Play testing feedback — see the note below the table | | `get_reviews_summary` | review count + average rating of one collection (absent `ext_id` = its own apps, as in `get_reviews`), each with `previous` (the preceding period of equal length) and `diff` — `apps_id` required, `ext_id` narrows to one app; optional `country`, `lang`, `from`, `to` (default the last 30 days); `total.value` sizes a `get_reviews` pull (review pages carry no total) | | `get_feedback_hub_feed` | paged Feedback Hub items (Steam reviews and discussions, subreddits, custom and other entry-point sources) of one collection or one app in it — `apps_id`, `from`, `to` required (`to` inclusive); optional `ext_id`, `q`, `lang`, `min_rating`/`max_rating`, `sentiment`, `semantic_tag_ids`, `tag_ids`, `extra_filters` (up to 10 `{path, op, value}` on `extra` fields), `shape` (`flat` default, `tree` with `children_limit`), `sort`, `limit` (1-100, default 50), `target_lang` (a BCP 47 tag for translated content); page by passing `page.cursor` back as `cursor` until it is null; store reviews stay on `get_reviews`; a collection without Feedback Hub sources returns an empty page that is still billed | | `get_feedback_hub_extra_keys` | the `extra` key paths (with JSON types) the Feedback Hub items of a collection carry — `apps_id`, `from`, `to` required, optional `ext_id`; run it before an `extra_filters` filter | | `get_reviews_ai_summary` | AI bullet summary of reviews | | `reviews_semantic` | per-review tags + sentiment | | `reply_to_review` | reply to a review — `apps_id`, `store`, `ext_id`, `review_id`, `answer_text`: `review_id` is the review's `reviewId` from `get_reviews`, `store` its `metaInformation.store` (preview→confirm) | | `reply_to_review_preview` | mint a confirmation token for `reply_to_review` | | `report_review_concern` | flag a review (preview→confirm) | | `report_review_concern_preview` | mint a confirmation token for `report_review_concern` | | `import_reviews` | import reviews into an existing source of the collection (preview→confirm) — `target` `custom_source` (default, the classic Reviews feed) or `feedback_hub` (the Feedback Hub, needs `platform_slug`); `ext_id` is the source app's `extId`, which must already exist; each review needs `ext_review_id`, `content_raw`, `ext_created_at` | | `import_reviews_preview` | mint a confirmation token for `import_reviews` | `get_reviews` takes an optional boolean `testing_feedback` (default `false`, Google Play only): - For a Google Play app, `true` returns ONLY testing-feedback reviews — what testers left on a closed or open testing track, each arriving with `typeList: ["beta"]` — and `false` or absent returns ONLY public reviews. The two sets are disjoint: for that app one call never returns both, so reading both takes two calls. - Other stores ignore the flag and always return their public reviews, so a workspace-wide `true` call mixes them in. Scope `true` calls with a Google Play `ext_id`. - Requires the Google Play testing feedback plan feature and a Google Play Console connection with Testing feedback access. Without the connection the set is empty — no error. - Same price per page either way (`references/costs.md`). `get_reviews_summary` does not take the flag: it counts public reviews only, so it never sizes a `true` pass. ## Ratings | Tool | Purpose | |---|---| | `get_ratings_history` | public ratings series of one app — `ext_id` + `store` (short code) + `from` + `to`; one `country` per call (default `all`, worldwide); `period`, `type` (`total` cumulative or `diff` incremental — no `diff` for `gp`); the full series in one call, no paging | | `get_gp_console_ratings` | Google Play Console ratings — `apps_id` + `ext_id` + `from` + `to` + `fields` (exactly one of `ratings`, `users`) | `get_reviews`, `get_reviews_summary`, `get_ratings_history` and `get_gp_console_ratings` take `YYYY-MM-DD` dates with `from` before `to` (`from` equal to `to` is refused). An `ext_id` on `get_reviews` or `get_reviews_summary`, and every `get_gp_console_ratings` call, first looks the app up in the collection. The lookup is free, and an `ext_id` that is not in that `apps_id` is `NOT_FOUND` before any data call. ## ASO | Tool | Purpose | |---|---| | `get_rankings` | category rank | | `get_keywords` | search downloads of one app's tracked keywords — per keyword its share of the app's search downloads and the download count (null without a store connection), each with its change over the period; one collection, country and device per call — `apps_id`, `ext_id`, `country`, `columns` (`["searchDownloads"]`) required; optional `store`, `device`, `from` + `to` (both or neither; default the 14 days ending yesterday), `sort`, `page`, `limit` (1–100 per page); ranks and scores are on `list_keywords` | | `list_keywords` | tracked keyword positions | | `aso_search_results` | SERP results / own card | | `aso_search_suggestions` | keyword suggestions | | `aso_search_ads_recommendations` | Search Ads recommendations (App Store) | | `aso_featured_apps` | deprecated: every call answers `TOOL_DEPRECATED`; use `get_app_featuring` | | `get_listing_changes` | store listing changes of tracked apps over a period: `apps` (`ext_id` + `store`), `from`, `to`, optional `countries`, `fields`, `page` | | `get_in_app_events` | in-app events of tracked apps over a period: `apps`, `from`, `to`, optional `countries`, `page` | | `get_app_featuring` | store featuring of tracked apps over a period (App Store and Google Play): `apps`, `from`, `to`, optional `countries`, `page` | | `get_aso_report_app_store` | App Store ASO report | | `get_aso_report_gdc_google_play` | Google Play ASO report | | `edit_keywords` | replace the tracked keyword set for one collection/country/device — keywords not passed are dropped (preview→confirm) | | `edit_keywords_preview` | mint a confirmation token for `edit_keywords` | ## Feedback | Tool | Purpose | |---|---| | `submit_feedback` | send a 1–4 session-satisfaction score + optional comment(s) back to AppFollow; optional opt-in attachments — see `references/feedback.md` |