AniList MCP Server (Unofficial)

[Requires login] Add an anime/manga to the authenticated user's own AniList list. Use search_media first to resolve the title to its AniList media ID. Only set the fields you care about — everything else is left at AniList's defaults ONLY if this media isn't already on the list. AniList upserts by media ID: if an entry for it already exists (e.g. previously dropped, with its own score/notes/progress), calling this updates that entry in place instead — every field you don't set here keeps its previous value, not a default. Use get_user_list first to check for an existing entry if you need a guaranteed-fresh one. The response echoes what AniList actually stored, so check it rather than assuming the write landed verbatim. It includes `customLists`/`advancedScores` only when you set them — those are the fields AniList zeroes/replaces rather than merges, so they're worth reading back; otherwise the echo is trimmed, since a list sync is hundreds of writes.

[Requires login] Delete an activity post the authenticated user owns. This cannot be undone, and calling it again on an already-deleted id errors as not found rather than silently succeeding.

[Requires login] Delete a forum thread the authenticated user owns, by its ID (from search_thread, get_thread, or the id returned by post_thread). This cannot be undone, and calling it again on an already-deleted id errors as not found rather than silently succeeding.

[Requires login] Delete a comment the authenticated user owns, by its ID (from get_thread_comments or the id returned by post_thread_comment). This cannot be undone, and calling it again on an already-deleted id errors as not found rather than silently succeeding.

Get a single AniList activity post (list update, text post, or message) by its ID, including `replyCount`, `likeCount`, and `isLiked`.

Get upcoming (or a specific title's) episode air times, soonest-airing first (confirmed live) — or, with `notYetAired: false`, already-aired ones newest first. Omit `mediaId` for the site-wide upcoming schedule, or pass it (from search_media/get_media) to get one title's next-episode air time. Set `airingAfter`/`airingBefore` for a date window instead (e.g. what airs this week, or what aired yesterday). Anime only — manga has no airing schedule; a manga id (like any non-anime id) is rejected as not-found rather than silently returning an empty schedule, though that error reads the same as a nonexistent id and doesn't single out the manga/anime mismatch.

[Requires login] Get the profile of the AniList account currently authorized via login_anilist/ANILIST_ACCESS_TOKEN — use this to confirm which account is connected. Also the authoritative source to fetch BEFORE an update_user call touching `options.notificationOptions`/`options.disabledListActivity` (confirmed live: full array replace, not merge by entry) or `mediaListOptions.animeList/mangaList.customLists` (confirmed live: same full-replace behavior) — read the current values here first and resend them in full alongside your changes. The sibling `advancedScoring` array carries a related but distinct risk instead: it's positional, not name-matched, so reordering/renaming its entries silently reinterprets already-scored list entries (see update_user's own `advancedScoring` field for detail) — fetch it here first too before changing it.

Get a character's profile by AniList character ID: name, image, description, and the anime/manga they appear in (`media`, most popular first, with their `characterRole` — MAIN/SUPPORTING/BACKGROUND — in each). Use search_character first to resolve a name to its ID. `media` is paginated (`page`/`perPage`, max 25): a long-running franchise character can appear in far more titles than one page holds, so check `media.pageInfo.hasNextPage` before treating it as the full list. Narrow it with `type` (anime or manga only) or `onList`.

Get an AniList user's profile AND statistics in a single call — use this instead of calling get_user_profile and get_user_stats separately. Like get_user_profile, this also returns the target's account settings (notifications, list display, etc.) regardless of who's authenticated — AniList doesn't viewer-gate those fields. Accepts an exact AniList username directly — no need to call search_user first unless you only have a partial/fuzzy name. Its embedded statistics carry the same staleness caveat as get_user_stats: AniList's own aggregation can lag and read all-zero even for an account with real list entries.

List every genre name AniList uses to tag anime/manga (e.g. Action, Comedy, Slice of Life). Call this before filtering search_media by genre, so you pass a name AniList actually recognizes.

Get information about one or more anime or manga by their AniList ID(s). At the default `format: "full"` that means the whole card: title, format, status, episode/chapter/volume count, genres, score, synopsis, dates, and `rankings` — AniList's own ranking badges (e.g. "#134 highest rated all time", "#11 highest rated 2024"), one entry per rated/popular ranking window the title currently appears in. Also returns `nextAiringEpisode` (for currently-releasing anime), `externalLinks` (official sites, streaming platforms), and — [requires login] — `mediaListEntry`, the authenticated user's own list entry for this title, or null if it isn't on their list. Identify titles by AniList ID (`ids`) or by MyAnimeList ID (`malIds`) — exactly one of the two, and either accepts a batch. Use search_media first only when you have neither: a title search costs a call per title and can match the wrong entry, so prefer `malIds` whenever you already have MAL IDs. For a batch whose point is mapping IDs to titles rather than reading them, pass `format: "compact"` instead, which drops everything above except the identifiers, titles, type/format/status and episode/chapter counts — measured at 5KB against 125KB for the same 30-title batch. Returns a single object if you passed a single ID, or an array (same order as the IDs you passed, with `null` in place of any that didn't resolve to a real anime/manga) if you passed an array.

List an anime/manga's characters with their role (MAIN/SUPPORTING/BACKGROUND) and, for anime, their voice actors. Use search_media first to resolve a title to its AniList ID. Grouped by role — MAIN first, then SUPPORTING, then BACKGROUND (confirmed live) — and in AniList's own relevance order within each. Voice actors come back in every dub language unless you set `voiceActorLanguage`; set `role` to list only one role (e.g. just the MAIN cast).

Get the anime/manga related to a given title, each with its `relationType`: PREQUEL, SEQUEL, PARENT, SIDE_STORY, SPIN_OFF, ALTERNATIVE, SUMMARY, CHARACTER, ADAPTATION (this title adapted into another format), SOURCE (the material this title was adapted from), COMPILATION, CONTAINS, SAME_UNIVERSE or OTHER. Use search_media first to resolve a title to its AniList ID.

List user-written reviews for an anime or manga, most-helpful-voted first (confirmed live: ordered by `rating`, not `score` — see those fields' own descriptions). Always includes `summary` (a short excerpt); set `includeBody` to also fetch each review's full text (can be long — leave it off unless you actually need the full text). Use search_media first to resolve a title to its AniList ID.

List an anime/manga's staff (director, writer, character designer, author, illustrator, etc.) with their role. Use search_media first to resolve a title to its AniList ID. No explicit ordering is requested — don't assume results are sorted by role or popularity (unlike get_staff's reverse-direction lookup, which IS popularity-sorted).

Get an anime/manga's watch/read-status counts (watching/completed/planning/etc.) and score distribution histogram across all AniList users. Use search_media first to resolve a title to its AniList ID.

List every descriptive tag AniList uses on anime/manga (finer-grained than genres, e.g. 'Time Skip', 'Tragedy', 'Reincarnation'), with category and adult-content flag. There are ~425 tags total, so results are paginated — use `page`/`perPage` rather than expecting them all in one response. Use this to look up a tag's exact name before passing it to search_media's `tag_in` — tag names are case-sensitive, and an unrecognized name doesn't error, it just silently matches nothing (confirmed live).

[Requires login] Get the authenticated user's AniList notifications: new episodes airing, activity likes/replies/mentions, new followers, thread replies/likes, and media/staff/character data-submission updates. Every item has `id`, `type` and a human-readable `context`/`contexts` string; the rest of the fields depend on `type` (e.g. an AIRING item includes `media`/`episode`, a FOLLOWING item includes `user`, an ACTIVITY_MESSAGE item includes `message.message` — the actual DM text is nested one level inside `message`).

Get a single AniList recommendation pairing (media + the media users recommend alongside it) by its AniList recommendation ID — useful for re-checking one specific pairing's current `rating`/`userRating` (e.g. after voting on it directly on anilist.co — this server has no tool to cast that vote itself) without re-paginating the whole list. Use get_recommendations_for_media instead to discover recommendation IDs for a title in the first place, or to browse/list them. Confirmed live (raw GraphQL, independent of this server): this is an AniList-side inconsistency, not a client bug — a majority of ids from get_recommendations_for_media's own `nodes[].id` 404 here anyway (7 of 10 tested for one title), with no correlation to that pairing's `rating`. A 404 on an id you just got from that tool doesn't mean the pairing doesn't exist, only that this specific lookup can't resolve it.

List anime/manga AniList users recommend as similar to a given title, ranked by rating (highest first). Use search_media first to resolve the title to its AniList ID. Each result's `mediaRecommendation.mediaListEntry` (requires login) shows whether it's already on your own list — set `excludeInList: true` to drop those from the fetched page instead of checking each one yourself (a local post-fetch filter: AniList's own `onList` filter applies to the source title, not the recommended one).

Get AniList's own site-wide statistics (new users/anime/manga daily counts), newest first. Useful for questions about AniList's growth/activity, not for anime data. Defaults to the last 7 days; use `page`/`perPage` to go further back. Confirmed live (bypassing this tool's own 25 cap): AniList's SiteStatistics field silently caps `perPage` at 25 itself, so this tool's own limit doesn't lose you anything.

Get a staff member's profile by AniList staff ID: name, image, occupations, description, and the anime/manga they hold production credits on (`staffMedia`, most popular first, with their `staffRole` — e.g. Director, Character Design, Theme Song Performance — in each). Use search_staff first to resolve a name to its ID. Voice-acting roles are NOT part of `staffMedia` (confirmed live: Kana Hanazawa's lists only her theme/insert song credits) — for a voice actor, pass `credits: "voice"` to get `characterMedia` instead: each title with the character(s) voiced and their `characterRole`. Either list is paginated (`page`/`perPage`, max 25) — check its `pageInfo.hasNextPage` before treating it as the full list. Narrow it with `onList`, or `type` (anime or manga only; production credits only).

Get a studio's profile (name, whether it's an animation studio) and the titles it's credited on (`media`, most popular first), by AniList studio ID or by name. `media` is paginated (`page`/`perPage`, max 25) — check `media.pageInfo.hasNextPage` before treating it as the studio's full catalog. By default it mixes titles the studio animated with ones it's only credited on (e.g. as a producer); set `isMain` to split them. Confirmed live: `name` already does AniList's own fuzzy search (same mechanism as search_studio), so a partial name (e.g. "Kyoto Anim") resolves directly — reach for search_studio instead only when you need to browse multiple candidates rather than take the closest match. If both `id` and `name` are given, `id` takes precedence and `name` is ignored.

Get an AniList forum thread's title, body and metadata (including `replyCount`, `viewCount`, `likeCount`, `isLiked`) by its ID.

List top-level comments posted on an AniList forum thread, by the thread's ID. Replies (posted via post_thread_comment's `parentCommentId`) are nested under their parent's `childComments` rather than appearing as separate entries in this list.

List AniList characters or staff members whose birthday (month/day) is today, most favourited first. Paginated — a single day can have well over a thousand entries, so check `hasNextPage` before treating a page as the full list. Returns character-shaped or staff-shaped objects for kind: CHARACTER/STAFF respectively, but a lighter fetch than get_character/get_staff: it does NOT include those tools' `media`/`staffMedia` filmography, nor each entry's bio (`description`) unless you set `includeDescription` — call get_character/get_staff by ID for an entry's full roles/works and bio, don't read a missing filmography here as 'none'.

List recent AniList activity posts from a specific user, newest first — by default every activity type (list updates, text posts, and message activity they received). Narrow it server-side with `type` and/or a `createdAfter`/`createdBefore` window rather than paging through everything and discarding. Accepts an exact AniList username directly (it's resolved to an id with one extra internal lookup) — no need to call search_user first unless you only have a partial/fuzzy name. Use search_activity instead to browse without pinning to one user. Use get_user_recent_activity instead if you just want a quick, fixed-size (5-item) recent snapshot without paginating.

Get a user's AniList anime or manga list. Returns a compact table of one row per entry by default (`format: "compact"`) — entry id, media id, MAL id, status, score, progress and title, which is what a sync, comparison or pick-something-to-watch task needs; switch to `format: "full"` only when you actually need dates, notes, custom lists, advanced scores or the status/custom-list grouping, since that response is roughly ten times larger per entry. Filter server-side with `statuses` rather than fetching everything and discarding — asking for just CURRENT costs a fraction of a whole-list read. Pass `sort` (e.g. UPDATED_TIME_DESC) to have AniList order the list before chunking, so a small `perChunk` returns the top entries by that key rather than an arbitrary slice. To answer "what did they finish/start in a period", use the `completedAt_*`/`startedAt_*` bounds (strict, server-side) instead of scanning the list — e.g. everything completed in 2026 is `completedAt_greater: {year: 2025, month: 12, day: 31}` plus `completedAt_lesser: {year: 2027}`; chunking then counts only the matching entries. Works for any public/unlisted user. Accepts an exact AniList username directly — no need to call search_user first unless you only have a partial/fuzzy name. Paginated by `chunk`/`perChunk` (AniList's own mechanism for this — counted across entries of ALL statuses combined, not per status), since a large list can otherwise return thousands of entries in one response; check `hasNextChunk` and increment `chunk` to keep paging. `perChunk` itself has no cap — AniList accepts values in the thousands, so raise it to pull a whole list in one call instead of paging through it. The collection AniList exposes here is capped at its ~11,000 most recently updated unique entries — irrelevant for virtually every account, but a hard ceiling if one is ever hit (older entries beyond it aren't retrievable through this field at all, by any chunk).

Get an AniList user's public profile: name, about text, avatar, donator status — plus their account settings (title/name-language preferences, notification toggles, list display options). Confirmed live these settings are NOT viewer-gated, so they're included for any user, not just the authenticated caller. Accepts an exact AniList username directly — no need to call search_user first unless you only have a partial/fuzzy name and need to look up the exact one. Need statistics too? Use get_full_user_info instead of also calling get_user_stats separately.

Get an AniList user's 5 most recent activity posts, with full content (not just IDs) — every activity type (list updates, text posts, and message activity they received; the underlying query applies no type filter of its own), a fixed count, not configurable; use get_user_activity for a paginated full feed instead. Accepts an exact AniList username directly — no need to call search_user first unless you only have a partial/fuzzy name (username resolution costs one extra internal lookup either way).

Get an AniList user's anime/manga statistics: counts, mean score, time watched, episodes/chapters/volumes consumed. Accepts an exact AniList username directly — no need to call search_user first unless you only have a partial/fuzzy name. AniList's own stats aggregation can lag behind the account's real list — confirmed live, an account with real scored/progressed entries still read back all-zero statistics — so don't treat a zeroed result as 'this account has no list activity' without cross-checking get_user_list. Need the profile too? Use get_full_user_info instead of also calling get_user_profile separately.

Authorize the personal-list and social tools with your AniList account (one-time). Prerequisite: register an app at anilist.co/settings/developer with Redirect URL set to `http://localhost:8082/callback` (the default; override the port via ANILIST_OAUTH_PORT if 8082 is taken — it must match exactly what you register), and set ANILIST_CLIENT_ID + ANILIST_CLIENT_SECRET in the server env. Calling this returns an authorization URL: open it, log in, and click Approve. If your browser genuinely runs on the same machine as this server, login then completes automatically; otherwise copy the URL you land on and pass it to submit_anilist_redirect. `auto_capture` in the response is NOT a reliable detector of that — it only reflects whether this server process could bind its own localhost listener, which typically succeeds even when the server runs remotely (SSH/headless) and the browser can never reach it; you (or the user) know the actual machine layout, the response doesn't.

[Requires login] Post a new message-style activity to another AniList user's profile, or update an existing one by passing its `id`. Despite the name, this is NOT a private DM — like post_text_activity, it's publicly visible on both users' activity feeds/profiles to anyone who can view them. This is a limitation of this tool, not AniList itself: AniList's underlying mutation does accept a `private` argument, but this tool doesn't expose it, so every message sent through it is public. Don't use it for anything the sender expects to stay confidential.

[Requires login] Post a new text-status update to the authenticated user's own AniList profile, or update an existing one by passing its `id`.

[Requires login] Post a new forum thread to the authenticated user's own AniList account, or update an existing one by passing its `id`. Use search_thread first if you want to check whether a similar thread already exists before posting a new one. Note: this tool's own response doesn't include `categories`/`mediaCategories`/`isSticky`/`isLocked` — call get_thread with the returned `id` afterward to confirm whether a category/sticky/locked change actually applied, especially since sticky/locked can silently no-op and categories/mediaCategories can silently drop existing values (see those fields' own descriptions below).

[Requires login] Post a new comment on an AniList forum thread from the authenticated user's account, or update an existing one by passing its `id`.

[Requires login] Delete an entry from the authenticated user's own AniList list by its list-entry ID (NOT the media id) — get it from get_user_list. This cannot be undone, and calling it again on an already-deleted id errors as not found rather than silently succeeding.

Search/browse AniList's activity feed (list updates, text posts, messages), optionally filtered to one user or — with `isFollowing` [Requires login] — to your following feed, to one or more activity types, and/or to a creation-time window. Each result already includes its full content (text/message/status, `replyCount`/`likeCount`/`isLiked`, the posting user, and — for list activity — the media) — no follow-up get_activity call needed unless you need to re-check one specific result's `replyCount`/`likeCount`/`isLiked` later (e.g. after liking it).

Search AniList for characters by name. Returns AniList IDs to use with get_character. Each result's bio (`description`) is omitted unless you set `includeDescription`.

Search AniList for anime or manga by title and/or filters (genre, tag, format, status, season/year, country, length — most with an exclusion variant), or browse term-less by ranking (top-rated, most popular, trending — see the `sort` parameter) when you don't have a title to search for. Adult (NSFW) results are included unless you set `sfw: true` (or `isAdult: false`); `isAdult: true` returns adult titles only. Returns AniList IDs to use with get_media and other ID-based tools.

Search AniList for staff members by name. Returns AniList IDs to use with get_staff. Each result's bio (`description`) is omitted unless you set `includeDescription`.

Search/browse AniList's animation/production studios by name, when you want to see multiple candidates rather than take the closest match — get_studio's own `name` param already does this same fuzzy search directly and resolves a partial name on its own. Returns AniList IDs to use with get_studio.

Search/browse AniList's forum threads by title/body text and/or restrict to one category. Returns AniList thread IDs (with title/category for context) to use with get_thread, get_thread_comments, post_thread, or post_thread_comment.

Search AniList for users by username. Returns AniList user IDs to use with get_user_profile/get_full_user_info (profile), get_user_stats (statistics), get_user_list (their public list), or toggle_follow_user.

Complete a login started with login_anilist by submitting the URL your browser was redirected to after clicking Approve (the one containing ?code=...). Use this when login didn't complete automatically — e.g. the server runs on a remote/headless host. Pass the whole URL, not just the code: its `state` is checked against the login attempt, so a bare code or a URL from another login is rejected. Fails if no login is in progress (e.g. the server restarted since login_anilist) — run login_anilist again.

[Requires login] Toggle an anime, manga, character, staff member, or studio in the authenticated user's AniList favourites. Calling it again on the same `kind`+`id` un-favourites it. Read `isFavourite` for the outcome: true = now favourited, false = now removed, null = can't tell (the list runs past the first 50 returned, and a newly added favourite goes to its end). `favourites` holds the first page (up to 50, id-only) of the toggled kind, under its key (anime/manga/characters/staff/studios). Resolve `id` first via search_media/get_media (kind: ANIME/MANGA), search_character/get_character, search_staff/get_staff, or search_studio/get_studio, matching `kind`. Confirmed live: AniList does NOT validate that `id` actually belongs to the given `kind` — e.g. passing an anime's ID with `kind: CHARACTER` succeeds silently instead of erroring, favouriting a nonexistent character. Always resolve `id` from the tool matching `kind` rather than reusing an ID you already have on hand. Note: immediately re-checking with get_media/get_character/get_staff/get_studio's own `isFavourite` can briefly still show the pre-toggle value — a confirmed AniList-side read-after-write lag, not a bug in this call; this tool's own `isFavourite` already reflects the new state.

[Requires login] Toggle following another AniList user from the authenticated user's account. Calling it again on the same user unfollows them.

[Requires login] Apply ONE set of values to many entries on the authenticated user's own AniList list in a single request — e.g. mark thirty titles COMPLETED, or set the same score on a batch. Use this instead of calling update_list_entry in a loop whenever the entries share the values you're writing; use update_list_entry when each entry needs DIFFERENT values (this tool cannot vary them per entry), or when you need to set `customLists`/`advancedScores`, which this tool deliberately doesn't accept. Takes list-ENTRY ids (not media ids) — get them from get_user_list. The whole call is all-or-nothing: if any id is unknown or isn't yours, AniList rejects the request and applies NOTHING, so there is never a half-applied batch to clean up — but the error does not say which id was at fault, so re-check the ids against get_user_list. Returns a summary (how many were updated, and their ids), not the full entries.