create_font_list
ChatGPTCreate a NEW font list (the API's FontSet — users say 'font list', 'list', 'font set' or 'font collection') in the user's Monotype library and add the given fontId(s) to it. Use it when the user wants to GROUP, SAVE or ORGANIZE fonts under a name. A font list publishes NO CSS and NO CDN URL, so it cannot license fonts for HTML/CSS/React or any rendered design — for that use create_web_project instead; if the request is ambiguous, ask which the user wants. Resolve font names to fontIds with font_search_by_name first. NAME CHECK before creating: call get_web_projects_by_name with the proposed name and assetTypes=["FontSet"]; on an EXACT name match the user is told the name is taken and picks another. FONT ELIGIBILITY: a .woff or .woff2 file is not eligible for a font list, whatever its source — WOFF is a web-delivery container and a list is for organizing/downloading font files. Check each font's fileName from the font_search_by_name result and filter those out; this is the mirror of the web-project rule, which REQUIRES .woff/.woff2 for imported fonts. Propose a name for the user to confirm; the tool does not check for an existing list of the same name. Production status does not gate list creation. Flow: POST /FontSet → PUT /FontSet/{assetId}/fonts → GET /FontSet/{assetId}. Adds any number of fonts — the upstream cap of 10 per call is handled automatically by batching in groups of 10. Returns the list's assetId and membership, with cssAvailable:false.
create_web_project
ChatGPTCreate a NEW Web Project (default) or a Digital Ad, add the given fontId(s), and return its published CSS / CDN URL. Pass assetType='DigitalAd' only when the user explicitly asks for a digital ad; otherwise it creates a web project. Takes fontIds from font_search_by_name. Propose a suggested name and let the user confirm/change it. NAME CHECK before creating: call get_web_projects_by_name with the proposed name and assetTypes=[the single type you are creating]; on an EXACT name match the user is told the name is taken and picks another, or reuses the existing asset. FONT ELIGIBILITY: every font whose source is "Monotype Fonts" is eligible, but an IMPORTED font (any other source) is eligible ONLY if its fileName ends in .woff or .woff2. Both source and fileName come from the font_search_by_name result. Before creating, check each font's production status via get_production_fonts_with_filters and show it to the user in a table (Font | In Production) — informational and non-blocking; if any font isn't production-marked, get the user's consent before creating. Flow: POST /{type} → PUT /{type}/{assetId}/fonts → GET /{type}/{assetId}. Adds any number of fonts — the upstream cap of 10 per call is handled automatically by batching in groups of 10.
font_search
ChatGPTLook up a font's detailed metadata by its fontId (MD5 hash). Uses POST `/v1/fonts/search` (Search & Discovery API). Pass the fontId returned by font_search_by_name or get_asset_details. Returns font metrics, classification, and catalog metadata. Example: {"fontId": "94d1e67a905d31a67124efb226807b2d"}. The fontId ↔ styleId ↔ psName mapping and the catalog metadata this returns are stable identity, so a result from earlier in the conversation may be reused.
font_search_by_name
ChatGPTSearch by font NAME across BOTH the Monotype Fonts catalog AND the customer's imported (custom-uploaded) fonts, returning matches with their fontId (MD5), styleId, source, fileName and metadata. Use it to turn a font name into fontIds — e.g. 'search <font> in my inventory', 'in my library', 'do I have <font>'; get_production_fonts_with_filters cannot resolve a name, and font_search takes only a fontId. A match proves the font exists, not that it is licensed or in production: for production status pass the fontIds on to get_production_fonts_with_filters as fontIds; there is no licensing check. Searches in tiers, each step running only if the one before found nothing: exact on name, then exact on psName (so PostScript names like AvenirNext-Medium resolve precisely), then a partial name search that widens to family level (AvenirNext-Medium -> Avenir Next Medium -> Avenir Next -> Avenir). The response reports which step won in matchLevel/searchedField/matchType, with exactMatch true when the row is the exact font asked for and fallbackApplied true when results were widened to the family. Calls POST /v1/fonts/search. Page size is capped at 50 and, by default, it paginates through ALL pages so a large family (e.g. 40+ styles) is returned in full — results are not truncated to one page. Response includes count, total, and pagesFetched.
get_asset_details
ChatGPTRetrieve detailed information about a single font asset by its type and ID. This is a direct lookup — not a tree traversal. Returns asset metadata including name, description, access types, size, published CSS path, creation date, and child font details with font family, style, weight, stretch, and subset information. Published CSS paths and child fonts can change, so re-read an asset before you emit its CSS or CDN URL. For anything else, a result from earlier in this conversation is fine.
get_production_fonts_with_filters
ChatGPTReports production status only; it is not a font search. A font name, including 'do I have / search <font name> in my inventory or library', resolves with font_search_by_name, whose fontIds are then passed here as fontIds. searchTerm is a loose substring match, so its empty result does not mean the user lacks the font. Retrieve the user's production fonts via the Monotype License Management API. Filters: searchTerm (name) or fontIds; no other filters are available. Each font row carries name, family, fontId, psName, inProduction, productionFontCount, lastActivityDate, acceptedBy and usedBy; a row found by fontIds carries the fontId you passed. A result from earlier in this conversation may be reused. Call again only when the user asks for current status, or when you are about to create or modify an asset that depends on it — not on every prompt. PRESENTATION: When this tool's result answers an analysis, report or summary request, it suits a visual dashboard — KPI tiles, sortable tables, a 'who triggered it' leaderboard, and a trend chart — or Markdown tables in text-only clients, rather than plain prose or a raw JSON dump. A check of one or a few fonts is a short answer.
get_web_projects_by_name
ChatGPTList or search the user's Web Projects, Digital Ads, and Font Sets (which users call 'lists') from their licensed font library. Call this when the user asks to see, pull, browse, or list their library, web projects, digital ads, font sets / lists, font projects, or licensed fonts — e.g. 'pull my library', 'show my web projects', 'list my digital ads', 'show my lists', 'what's in my Brand list', 'what fonts do I have', 'list my projects'. A FontSet IS a 'list' — filter with assetTypes=["FontSet"] when the user asks about list(s). Searching by name (name) covers ALL THREE asset types and BOTH access types (owner and shared) by default, so leave assetTypes/accessTypes unset unless the user was specific. Listing without a name paginates at 100 per page: show the first page with the total and ASK the user whether to fetch the next one (call again with pageNumber incremented) rather than paging through automatically. Check hasNextPage / nextPageNumber in the response. Sort with sortBy (name | createdAt) and sortOrder (asc | desc). Returns lightweight metadata per asset: id, name, assetType, createdAt, parentId, accessTypes. It does NOT return child fonts or a child count — use get_asset_details with the returned id for publishedCSSPath, child fonts and size. A listing from earlier in this conversation may be reused. Call again when the user asks what exists now, or before creating or editing an asset by name — not on every prompt.
list_synced_fonts
ChatGPTList the fonts currently synced to the user's Monotype Fonts DESKTOP app, grouped by family. Use for 'what's synced', 'what fonts are on my desktop', and as the first step of any unsync request. Each family carries its familyId, and each file its variationId, name, psName, md5 and syncedOn — pass those ids to unsync_fonts. Present the result as a table (Family | Styles | Synced on). Read-only.
sync_fonts
ChatGPTSync fonts to the user's Monotype Fonts DESKTOP app, so they can use them in desktop applications (Word, Figma desktop, Adobe apps, ...). Use for 'sync / activate / install / add to my desktop' requests. Not for web use (that is a web project) or file download (that is font_asset_delivery). How to build the request: - A whole family ('sync Ringside', 'sync the complete Avenir family'): pass familyNames: ["Ringside"] — the server finds the graph family id and syncs the whole family with ONE syncFamily call, so there is no need to call font_search_by_name and pass every font. Only a family named EXACTLY as requested is synced; look-alikes are never substituted. If a name lands in unresolved with candidates (e.g. 'DIN Next J'), show them and ask which to sync, then pass the chosen one's familyId as familyIds. relatedFamilies on a match (e.g. 'Ringside Wide') are optional extras to offer. - Specific weights / styles ('sync Gotham Bold and Book Italic'): font_search_by_name for each, pick the exact matches, and pass their fontIds as fontIds (preferred: the server maps each file to its graph style) or their styleIds as styleIds (converted server-side to the graph's numeric style id). - Everything in a brand guideline: extract the font names from the document, resolve them all with font_search_by_name, then make ONE sync_fonts call with all their fontIds. - Everything in a web project / digital ad / font list: find it with get_web_projects_by_name, then pass assets: [{assetType, assetId}] — the server reads its fonts and syncs them all in one call. Several assets may go in one call. - You may also pass MD5 fontIds; they are resolved to styles server-side. CONSENT: syncing changes the user's desktop. When the request is broad (a whole family, a guideline, a project) or a font match was not exact, first call with dryRun: true, show the user what will be synced, and call again without dryRun after they confirm. Response: mode (success | partial | error | dry_run), message, affectedCount, families, variations [{variationId, name, psName, md5, familyName, syncedState, syncedOn}], mapping [{fontId, searchStyleId, graphStyleId, familyName, matchedOn}] (how each MD5 was matched 1:1 to a graph style), unresolved (items that could not be resolved, incl. ambiguous ones with candidates — name them to the user), errors (upstream failures, e.g. a font not licensed for desktop). Report from variations, not from the ids you sent. BATCHING: pass everything the user asked for in ONE call — there is no request limit. The server splits it into upstream calls of at most 5 families and 50 styles each (e.g. 12 families = 3 calls). operations shows the calls made (e.g. 'syncFamily x3'); a failed batch is listed in errors while the other batches still apply.
unsync_fonts
ChatGPTUnsync (remove) fonts from the user's Monotype Fonts DESKTOP app. The fonts stay in their library; only the desktop activation is removed. Use for 'unsync / deactivate / remove from my desktop' requests. Preferred flow: call list_synced_fonts first and pass the variationIds (single files), or familyIds (whole family) it returns; for 'unsync the Ringside family' you can pass familyNames instead — those ids are exactly what is synced. Name-based requests can also use styleIds from font_search_by_name, and 'unsync everything from project X' can use assets: [{assetType, assetId}]. CONSENT: removing fonts from the user's desktop apps can break open documents, so the user is shown exactly what will be removed (with dryRun: true when the request is broader than ids from list_synced_fonts) and confirms before the real call. Response mirrors sync_fonts: mode, message, affectedCount, families, variations, unresolved, errors. BATCHING: pass everything the user asked for in ONE call — there is no request limit. The server splits it into upstream calls of at most 5 families and 50 styles each (e.g. 12 families = 3 calls). operations shows the calls made (e.g. 'unSyncFamily x3'); a failed batch is listed in errors while the other batches still apply.
update_font_list
ChatGPTAdd font(s) to and/or remove font(s) from an EXISTING font list (the API's FontSet — users say 'font list', 'list', 'font set' or 'font collection') in the user's Monotype library. This is the right tool for 'add <font> to my <named> list' / 'remove <font> from <list>' requests; use create_font_list to make a new list, and update_web_project for a web project or digital ad. Resolve font names to fontIds with font_search_by_name first. The list is found by exact name (or pass assetId to skip the lookup); if no list matches, nothing is changed, and if several share the name the tool asks which. At least one of addFontIds / removeFontIds is required; adds are applied before removes. A font list publishes NO CSS and NO CDN URL, so it cannot license fonts for HTML/CSS/React or any rendered design. FONT ELIGIBILITY (adds only): a .woff or .woff2 file is not eligible for a font list, whatever its source — check each font's fileName from the font_search_by_name result and filter those out; this is the mirror of the web-project rule, which REQUIRES .woff/.woff2 for imported fonts. Production status does not gate list edits. Flow: resolve assetId → PUT /FontSet/{assetId}/fonts with action add and/or remove → GET /FontSet/{assetId}. Handles any number of fonts — the upstream cap of 10 per call is batched automatically. Returns the list's refreshed membership, with cssAvailable:false.
update_web_project
ChatGPTAdd font(s) to an EXISTING web project or digital ad (edit). This is the right tool for 'add <font> to <named/existing asset>' requests (e.g. 'add Gotham to check097'); the font name resolves to a fontId with font_search_by_name. FONT ELIGIBILITY: every font whose source is "Monotype Fonts" is eligible, but an IMPORTED font (any other source) is eligible ONLY if its fileName ends in .woff or .woff2. Use create_web_project to make a new one instead. Default assetType is WebProject; pass 'DigitalAd' for a digital ad. Adds any number of fonts — the upstream cap of 10 per call is handled automatically by batching in groups of 10. Resolves the assetId from its name, PUTs /{type}/{assetId}/fonts with the fontIds, and returns the refreshed published CSS / CDN.