MCP App Store
by KeenEthics
WeStudy icon

WeStudy MCP Server for ChatGPT

By WeStudy

Overview

WeStudy helps educators, course creators and school administrators create and organize online courses in the WeStudy learning management system (LMS). Prepare lessons and assignments, build question banks, and schedule when learning materials become available. Connect an existing WeStudy school to manage its teaching content through conversation, within the permissions granted to your connection. Create unpublished courses and update their settings, pricing, tariffs, calendar plans and material release schedules. Build course structure with sections, lessons and assignments; edit drafts and prepare private course video references. Publish or unpublish courses and materials when explicitly requested. Configure shared course styles or the style of one course. Course test materials, certificates, student enrollment and grading submissions are not included in this version. Create question banks, add supported question types in bounded batches, and read or edit existing questions. Add or update a contact, create or edit CRM custom fields, and configure the fields required for course registration. Search files and folders with server-side sorting and limits, rename files, create folders, move selected files, and import files from supported sources. Local file uploads require a client that can access local files and run the supplied upload helper. The plugin cannot permanently delete school files or folders. Create and edit draft pages, the homepage and course landing pages. Use the constructor and rich-editor reference catalogs, validate native JSON, and generate a temporary 15-minute design preview. Visual inspection requires browser tools in the client; without them, the plugin discloses that the design was not visually checked. Publication is separate from draft editing. Shared design settings and course settings take effect when changed. Operations may finish asynchronously. The plugin checks execution status and does not automatically repeat a write with an uncertain outcome. An existing WeStudy school and authorized connection are required. The plugin does not provide checkout or process payments.

How to connect

  1. ChatGPT: Open the WeStudy listing in the ChatGPT app directory and choose Connect. ChatGPT listing WeStudy exposes 81 tools here.

Tools

add_edit_contact

ChatGPT
Creates a contact in the connected school. If its email already exists, updates only supplied fields. Use only when the user requests this write. Never retry an UNKNOWN execution; use get_execution to check its state.

configure_course_student_import

ChatGPT
Configure a prepared Excel student import using the current import revision. Either choose zero-based sheet and one-based headerRow for another preview, OR map zero-based column indices to supported native field names/null (email required). Mapping validates rows in the existing worker; it does NOT enroll. Poll get_course_student_import until ready. List imports need no mapping.

create_course

ChatGPT
Create one inactive new-version course with settings; no materials, tariffs, calendars, students or landing page are created. The primary language defaults to school language. Initial accessType may be all or days_after_register; create a calendar plan before enabling scheduled through update_course. Single-price paid courses require price. Use publish_course separately only when requested. Preserve requestId and poll get_execution; never duplicate an uncertain create.

create_course_element

ChatGPT
Create an unpublished module, lesson or task in a new-version course. Read get_course_structure first and use expectedStructureRevision. Optional native page is saved only as draft. Task forms use isTaskReply=true and text_field/text_area/file_input questions, without form/question entity IDs. No tariffs, schedules, tests or students are changed. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

create_course_elements

ChatGPT
Create 1–100 EMPTY unpublished modules, lessons and tasks atomically in one existing new-version course. Prefer this over repeated create_course_element calls for outlines or adding a module with empty children. Read get_course_structure first for expectedStructureRevision. Each item has ref, type, name and optional parentUuid (existing module) OR parentRef (module earlier in this batch). Omit both for root. Parents precede children; append siblings in input order. Modules count toward the 100-item limit. Does not edit existing materials, fill content, publish, enroll students or change tariffs/schedules. Result maps refs to real UUIDs and returns structureRevision for the next sequential batch. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

create_course_group

ChatGPT
Create a course group using existing native group rules. Calendar courses require a verified scheduleUuid; other courses omit it. Does not enroll students.

create_course_schedule

ChatGPT
Read get_course_schedules first and use its list revision. Create one calendar plan with name, education start/end and optional registration dates/open flag. Dates require ISO 8601 with seconds and timezone offset. Registration defaults to education dates; enrollmentOpen defaults false. Does not enable scheduled access, assign material dates, publish or enroll students. Reread the list before another create. Preserve requestId and poll get_execution.

create_course_tariff

ChatGPT
Read get_course_tariffs first and use its list revision. Create one tariff with name, paid price or isFree=true, optional access days, visibility and certificate limit. Does not enable tariffs, assign materials or alter students. update_course controls isPaid/hasPaymentPlans. Reread list before another create. Preserve requestId; poll get_execution after uncertain outcomes.

create_course_test

ChatGPT
Create an EMPTY unpublished test material at root or inside an existing module of a new-version course. Read get_course_structure first for expectedStructureRevision. Test-specific settings are optional beyond name. Does not publish, assign tariffs or change other materials. Populate with update_course_test_questions or create_course_test_questions, then reread. QUEUED/RUNNING: poll get_execution. Preserve requestId after timeouts or UNKNOWN; never blindly repeat writes.

create_course_test_questions

ChatGPT
Create 1–50 complete bank questions AND attach them to this test in one transaction, preserving supplied order and points. Read get_course_test first. bankUuid selects a verified bank; explicit null uses/creates the test own bank in this course. Follow create_questions content rules. Save supplied content when authorized; agree on a generated set once. No publication. A failed batch leaves neither new questions nor attachments. Use returned UUIDs and verify result before continuing. QUEUED/RUNNING: poll get_execution. Preserve requestId after timeouts or UNKNOWN; never blindly repeat writes.

create_custom_field

ChatGPT
Create one school-wide CRM field. Read get_custom_fields first to avoid unintended duplicate fields. Supported types match the CRM editor; select needs text/value options. Does not attach to courses or populate contacts. Keep requestId and poll get_execution after uncertain outcomes.

create_file_uploads

ChatGPT
Prepare 1-10 direct chunk upload sessions for existing LOCAL files. Use ONLY when this environment can read those exact local files and execute the plugin local uploader. Ordinary ChatGPT without local access should recommend uploading large/bulk files into a folder through the WeStudy cabinet, then search_files. Inspect exact name/size/MIME first; no file bytes, base64 or local paths are sent to this tool. A local helper sends 50 MiB chunks straight to the returned uploadUrl/next_url, then confirms the storage receipt at completionUrl with completionToken. Keep tokens/URLs out of chat text. Preserve requestId and transferIds; no blind retry after a lost chunk response. Reuse existing completed files. Preparing a session is not a completed upload. No conversion or course attachment.

create_folder

ChatGPT
Create a school storage folder with name and optional parentUuid (null=root). Use one requestId for this exact operation; retry with the same requestId only. Same names may represent different folders; never silently merge. Resolve destination first. A clear request to create the folder authorizes this action.

create_page

ChatGPT
Create an unpublished standalone page, homepage or course landing, optionally with native design. Read get_site_config first for validation context. New form/question entity IDs must be omitted. Existing homepage/course custom pages cause CONFLICT; edit the returned existing page instead. Does not publish or save site styles. Metadata requires title/alias for standalone pages. Preserve requestId and poll get_execution; never duplicate an uncertain create.

create_page_translation

ChatGPT
Create an unpublished translation draft on the SAME existing page in an enabled school language. Read get_page_translations first and pass its revision. source=primary copies the primary draft (or saved content) with independent form/question identities; source=empty starts empty. Text is NOT translated automatically: translate the returned native page and save via update_page with its lang/revision, preserving all layout, IDs and non-text fields. An existing target is returned unchanged, never overwritten. Does not publish, change the primary translation, alias, school languages or site styles. Preserve requestId and poll get_execution; do not repeat an uncertain write.

create_question_bank

ChatGPT
Create one empty question bank in the connected school. Requires its name and an explicit target: courseUuid=null for a school-wide bank, or a real course UUID from search_courses / a verified prior result. If the user has not specified the target, ask whether the bank is school-wide or belongs to a course. Resolve ambiguous course matches with the user before writing; never choose the first match or invent a UUID. An explicit request to create with a resolved name and target authorizes this action. Returns the created bank UUID, name, zero questions and course details. Each call creates a NEW bank, even with an existing name. For QUEUED/RUNNING poll get_execution; never automatically retry a write after an error or uncertain result. Does not create questions, courses or tests.

create_questions

ChatGPT
Create 1–50 questions in one verified bank, atomically. Supports default (choices), match (pairs, all pairs required), inputAuto (case-insensitive exact text), inputManual (teacher grades). Preview generated content and obtain one approval for the complete set; an explicit request to save supplied content already authorizes it. Resolve the bank/course first. Split large sets into batches of at most 50 and 512 KiB including converted content. Reuse requestId only for identical retries. Poll get_execution for QUEUED/RUNNING; UNKNOWN requires verification, never a new requestId retry. Image blocks require a ready public file from this school: use import_files/get_file_transfers first for new photos. This command itself does not upload bytes.

create_site_section_translation

ChatGPT
Create and immediately publish a missing header or footer in an enabled school language. Read get_site_sections(lang=target) first for expectedRevision. source=empty starts empty; a source language copies its latest saved content with independent node/form identities and requires sourceRevision from reading that language. Text is NOT translated automatically. An existing target is returned unchanged. Never changes page content/publication, school languages, shared styles or other section translations. A live empty section replaces fallback in this language. Preserve requestId and poll get_execution; no blind retries.

get_course

ChatGPT
Read an existing new-version course in the connected school. Returns primary-language settings, publication status, currency, tariff flag, access mode and revision. Does not read students, materials, tariff entities or design and does not create anything. QUEUED/RUNNING: poll get_execution.

get_course_design_config

ChatGPT
Read active course-material styles for an explicit scope: parent=courses is the school-wide course default; parent=course requires courseUuid and returns its override or inherited courses/default config. source tells where styles come from. Reading never creates configs. Website and course landing styles use get_site_config instead. QUEUED/RUNNING: poll get_execution.

get_course_element

ChatGPT
Read the exact language material draft page, publishedPage, settings, current course generalConfig and revisions. Omitted lang uses course language; never silently edit another translation. Unsupported material types are readable but not editable. Reading never creates forms or certificates. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

get_course_groups

ChatGPT
Read course groups with schedule UUID, student count and revision. Groups in calendar courses belong to one schedule.

get_course_material_schedule

ChatGPT
Read material opening/closing boundaries for explicit mode=days or calendar. Calendar requires a verified scheduleUuid; days has none. Returns material hierarchy, names, isActive, timezone context and revision; no content or student data. Inactive mode may be read without enabling it. QUEUED/RUNNING: poll get_execution.

get_course_registration_fields

ChatGPT
Read this course registration field list, current definitions, required flags and revision. New-version courses in the connected school only. Does not read students or values. Global fields are available via get_custom_fields. QUEUED/RUNNING: poll get_execution.

get_course_schedules

ChatGPT
Read all calendar plans (including past/future), course accessType and timezone context. Each plan has an item revision for edits; the list revision is for creation. Does not activate schedules or read/enroll students. QUEUED/RUNNING: poll get_execution.

get_course_structure

ChatGPT
Read all course materials as a compact flat list with parentUuid, sibling sortOrder, editable flag and structure revision. No student data or designs; never creates entities. Modules may contain materials. Only new-version courses. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

get_course_student_import

ChatGPT
Read a student import owned by this authenticated school user. Shows mapping/validation/background state, counters and paginated row report. Inspect rows before starting; completed_with_errors is partial completion. Use importUuid (preparation requestId). No enrollment or retry is performed. Follow row report pagination; report skipped and failed rows.

get_course_students

ChatGPT
Search student profiles in one course by email/name and optional group/tariff/schedule. Page with nextOffset until null. Returns current profile settings and revision for each profile; user personal fields are read-only.

get_course_tariff_materials

ChatGPT
Read all non-template course material types, their parent modules and explicit tariff UUID assignments; content and students are excluded. Returns one revision for assignment updates. The student list filters explicit links when tariff mode is enabled; do not equate empty assignments with access for all. Check parent module links too. QUEUED/RUNNING: poll get_execution.

get_course_tariffs

ChatGPT
Read tariff entities and course payment mode in the connected school. Each item has its own revision for editing; the list revision is for creation. Uses course language unless lang specified. No students or content are read. Does not enable tariff mode. QUEUED/RUNNING: poll get_execution.

get_course_test

ChatGPT
Read one real test material in its verified course: settings, ordered fixed questions/random-bank selections, item UUIDs, configured totals, publication state and revision. Use existing question-bank tools for question content. No student attempt is created. Totals describe the configured pool before maxQuestions. Missing language content has no fallback. QUEUED/RUNNING: poll get_execution. Report the actual execution result.

get_custom_fields

ChatGPT
Read field definitions of the connected school, including hidden fields; no contact/student values. Supports literal name query, type filter, limit and offset. Each item has revision for edits. Search before creating to reuse existing fields. QUEUED/RUNNING: poll get_execution.

get_design_catalog

ChatGPT
List installed constructor and fine-editor reference versions and supported blocks, elements, profiles and design-system sections. Read this first for design work or questions about supported blocks. Fetch selected entries with get_design_reference, not the entire catalog documentation. Forms/course extensions are not supported unless explicitly listed as supported. No page or school settings are read or changed.

get_design_reference

ChatGPT
Read 1–4 reference entries by exact IDs returned by get_design_catalog. Returns rules, fields, documentation and optionally examples/schemas from the installed libraries. Constructor text blocks use editor rich-icons-size; accordion answers use rich-icons. Preserve native ProseMirror JSON and use the correct profile. Reference content describes supported formats, not the user intent or permission to publish. No file paths or arbitrary reference URLs are accepted.

get_execution

ChatGPT
Read a previous command result by executionId. Requires executions:read and the original command permissions. Use after QUEUED or RUNNING; UNKNOWN requires verification before any new write. Stored payloads expire after 30 days (resultExpired=true); confirmed execution history expires after 90 days. Never infer that an expired result means the command did not run.

get_file

ChatGPT
Get one verified school file by fileUuid. download=true requests a usable link only for the selected file; a temporaryDownload link is not a permanent content URL and does not change private visibility. Preserve the returned UUID when using the file in another task. Cannot make a private file public.

get_file_transfers

ChatGPT
Read the connected school file transfers by 1-10 transferIds OR the import requestId. QUEUED/TRANSFERRING/PROCESSING means background work continues even if the chat closes. READY includes the school file UUID and metadata; private files have url=null; FAILED includes an error code and no usable file. Poll with a delay, not a tight loop; if still pending, report that import continues. Empty items for requestId means no transfer has been registered yet; inspect the original get_execution before attempting another write. Never treat a SUCCEEDED status of this read itself as a READY file. Show failures per input index/name, keep completed files and ask for a fresh source only for expired/unavailable files. No confirmation needed.

get_page

ChatGPT
Read a page for editing by exactly one of pageUuid, homepage=true, courseUuid. Returns the exact language draft when present, otherwise published content/fallback, with contentLang and revision. source=default/empty and pageUuid=null means there is no custom page yet; use create_page. Preserve all native fields. For visual preview combine page with get_site_config. Read does not create anything. Webinar landings are not supported. QUEUED/RUNNING: poll get_execution.

get_page_translations

ChatGPT
Read the enabled school languages, primary language and existing translations of one custom page. Does not create translations or change school settings. Use exact language codes from languages; a missing entry in translations means that language version does not exist. Read its native draft/content with get_page(pageUuid, lang). To create a missing version use create_page_translation with this revision; never create a second page. QUEUED/RUNNING: poll get_execution.

get_pending_evaluations

ChatGPT
Read the pending teacher-evaluation queue counts for the connected school or one verified course. Omit courseUuid for the whole school; resolve a named course with search_courses first. Matches General evaluation: waiting flag and teacher evaluation type. pendingReplies is the primary queue count: student-material evaluation records, not individual form answers. pendingStudents counts distinct platform users across courses; pendingCourses counts courses with pending records. includeStudents=true optionally returns a bounded list with student UUID/name and their pendingReplies/pendingCourses, sorted by most pending replies first; limit defaults 20 (max 100), continue nextOffset only when more results are requested. Default returns counts only and students=null. One student may have several replies; never call pendingReplies the student count or sum per-course student counts as a school total. Names may be duplicated: use userUuid for identity. No answers, scores, contacts or files; no grading or writes. checkedAt is the UTC read time; the live queue can change between reads. QUEUED/RUNNING: poll get_execution.

get_profile

ChatGPT
Return the school represented by the authenticated connection. Use to identify the connected school. The profile ID stays the same after token refresh, reconnection, key replacement and school renaming.

get_question

ChatGPT
Read one verified question with its complete simplified content, answer/pair UUIDs, grading settings, languages, student-use flag and revision. Use this before update_question. Supported formatting is normalized; complex unsupported editor content requires editing on the platform. Existing image/video references are preserved; no media upload. This read needs no confirmation. For QUEUED/RUNNING poll get_execution.

get_site_config

ChatGPT
Read the active school SITE design system for page creation, editing and preview. Returns generalConfig and revision, including the default if no school config exists. Does not create or save a config. Course lesson design systems are separate and not handled by this tool. QUEUED/RUNNING: poll get_execution.

get_site_sections

ChatGPT
Read localized school header/menu and footer, enabled languages and per-section revisions. Missing translations expose fallback content and translationExists=false; create that language before editing. Reads do not write or publish. Header content excludes shared logo/styles; read get_site_config for those. Footer is one native constructor block. Saves through update_site_section publish only the selected section/language immediately, never pages. QUEUED/RUNNING: poll get_execution.

import_files

ChatGPT
Start background import of 1-10 files from HTTPS sources into the connected school storage. Supports documents, PDFs, video, audio, images and other files; preserves original bytes. No fixed 5 MiB limit or school tariff quota check; an explicitly configured storage file-size limit may apply. purpose=file defaults to private; purpose=editor_image requires a supported raster image and stores it publicly for editor content. Use real client file references in files OR ordinary HTTPS sources in urls, never both. The server accepts HTTPS file references, never raw local paths, sandbox URLs, base64 or invented links. A capable host may upload a local file using files automatically. For large local files use create_file_uploads and the local helper. folderUuid optionally selects a verified school folder, otherwise root. The source is temporary; never use it as a permanent content URL. A successful execution means transfers registered, not files ready. Poll get_execution for a queued tool execution, then get_file_transfers until READY or FAILED. READY.file.fileUuid identifies the saved file; url is null for private files. Only a READY public image URL can go in an image block. Importing does not attach a document to a course or convert video. Preserve requestId and recover via get_file_transfers(requestId) before retrying after an uncertain outcome. Report failures per file and reuse completed imports. The user request to store specified files in a resolved school authorizes this import

list_questions

ChatGPT
List or search question text in one verified question bank; optionally filter type. Returns UUID, type, preview and points, not full answers. Search scans stored text across languages. Continue nextCursor while hasMore=true, including empty filtered pages. Resolve ambiguous questions with the user, then get_question before editing. This read needs no confirmation. For QUEUED/RUNNING poll get_execution.

move_course_element

ChatGPT
Read get_course_structure and use expectedStructureRevision. Move one material with its entire subtree into a module or root (parentUuid=null), before a sibling or append. Other material content, tariffs, schedules and publications stay intact. No cross-course moves or cycles. Moving published content changes navigation immediately. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

move_files

ChatGPT
Move 1-100 verified files into folderUuid, or explicit null for root. One file uses a one-item array. Validation of the entire batch precedes any move. Returns moved and alreadyInFolder UUIDs. Bytes, URLs and visibility do not change. Resolve destination and sorting criteria before moving; do not infer document contents from filenames. Clear user selection and destination authorize the operation. For larger sets use consecutive batches and report partial progress. No folder moves or deletion.

prepare_course_element_file

ChatGPT
Use a READY private school file from import/upload/search, after creating the material identity. Add COURSE_ELEMENT scope without deleting existing scopes or changing visibility. kind=video creates/reuses a platform Video and returns video for config.video; kind=file only prepares a file scope; private document rendering/attachments are not supported in this stage. Use constructor video elements, not rich-text video_block, for private video. Does not insert into draft or publish: update_course_element separately. Never invent file/video UUIDs or expose signed private URLs in preview. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

prepare_course_student_import

ChatGPT
Prepare an XLS/XLSX file already READY in this school storage (fileUuid), up to 12 MiB and existing importer row limits. Upload privately with import_files/create_file_uploads first; never make student data public. Uses existing import queue; does NOT enroll yet. Poll get_course_student_import until mapping, configure columns, poll until ready, inspect errors, then start. Supports add_only/upsert and common defaults. No CSV, raw local paths or arbitrary URLs. Recover importUuid=requestId after uncertain preparation.

prepare_course_student_list

ChatGPT
Prepare 1-100 student rows from user text or supported image extraction through the existing import system. Does NOT enroll yet. Verify uncertain OCR emails; never invent addresses. add_only skips existing course profiles; upsert updates supplied fields. New users are matched/created within this school by email at start. New profiles default active and unpaid; tariff is independent of paid status. Uses the same background row processor as Excel. Inspect get_course_student_import then start_course_student_import. Row values override defaults. Reuse requestId on uncertain outcome and recover importUuid=requestId.

preview_design

ChatGPT
Validate {page, generalConfig} and create a temporary viewer URL through authorized PHP. The random link is readable by anyone who has it until its 15-minute expiry; it is not bound to a browser session. Does not create/publish a page or change the school design system. Fix validation errors first; use validate_design for details. For QUEUED/RUNNING poll get_execution instead of resubmitting. Expiry starts when PHP creates the cache; request a new preview when it expires. With a browser tool inspect the actual rendered design at desktop/mobile sizes; without one disclose that visual verification was not performed. For course materials pass context=course with the current course generalConfig; task reply forms use this context. General viewer does not verify student access to private video. Maximum design JSON 480 KiB and encoded PHP request 512 KiB.

publish_course

ChatGPT
Only for an explicit request to publish or activate this course. Use expectedRevision from the reviewed current course. Toggles course active=true like the platform publish button; does not publish material drafts or the landing page and does not enroll students or send notifications. Hidden/stopped flags are preserved. Preserve requestId and poll get_execution.

publish_course_element

ChatGPT
Explicit publication only. Read current material and use expectedRevision. Publish its selected-language draft through the platform form/assignment pipeline. Module publication here is NOT recursive. Does not activate the course, publish its landing, assign tariffs or notify/enroll students. Reread after a conflict; do not publish an unseen newer draft. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

publish_page

ChatGPT
Only for an explicit request to publish this page/these changes. Use pageUuid, lang and expectedRevision of the reviewed/saved draft. Calls platform publication: prepares form identities, copies draft to published content, clears the draft and enables the page. Does not modify site styles. Publication enables page visibility globally; previously published translations become accessible too. On CONFLICT reread/review; never publish unseen newer changes. Preserve requestId and poll get_execution.

rename_file

ChatGPT
Rename one verified file in the connected school. name is the new display name; if it has no extension the existing extension is retained. The file UUID, bytes and URL do not change. Resolve duplicate names with the user first. An explicit rename request with a resolved file and name authorizes this action. Verify uncertain results with get_file before retrying.

search_courses

ChatGPT
Search courses by a literal part of their name in the connected school. Returns candidate UUIDs, translated names, active flag, creation date and landing metadata (pageUuid, hasCustomPage, usesDefault, isPublished). For a course landing read get_page with the chosen courseUuid; create_page kind=course only if it has no custom page. If several courses fit the request, ask the user to choose; never silently select the first. Use the chosen UUID as courseUuid in search_question_banks. No confirmation is needed for this read. For QUEUED or RUNNING use get_execution.

search_files

ChatGPT
Search school files AND folders by literal name and MIME, or list folder contents with an empty query. location=root (default), all, or verified folder UUID. query always respects location. kind=files|folders|all. sortBy=size|name|dateAdded and sortOrder=asc|desc (default desc) sort on the server before limit/offset; omitted sortBy preserves ID-descending order. Mixed lists always keep folders first; folders support name ordering, otherwise keep their existing order. For top N largest files use kind=files, location=all (or the requested folder), sortBy=size, sortOrder=desc, limit=N (max 100); do not fetch every file to compare sizes. Smallest uses size asc, oldest/newest uses dateAdded asc/desc. Unknown sizes/dates sort last; equal values use ID descending. Stop once the requested N items are collected even if hasMore=true; follow nextOffset only when more requested results are needed, retaining filters and sorting. To find a folder by name use kind=folders and location=all, disambiguate using folderPath, then list its UUID. A top-N result is not the complete inventory; never derive school-wide totals or percentages from it. This does not search inside documents or analyze video content. No confirmation is required for this read.

search_pages

ChatGPT
Search stored pages by literal title/alias in the connected school, with bounded cursor pagination. Filter kind or courseUuid. Returns lightweight metadata, never designs. Default homepage/course templates are not stored school pages: resolve them with get_page. For course landing requests first use search_courses and its landing metadata. Do not silently pick ambiguous matches. QUEUED/RUNNING: poll get_execution.

search_question_banks

ChatGPT
Search question banks in the connected school. query filters bank names; optionally filter by courseName text in the same request, or by an already selected courseUuid. Never send both course filters. Returns bank UUIDs, translated names, question counts and course details; course=null means a school-wide bank. Text filters can match multiple courses and banks. Present ambiguous candidates for selection; a page is not the complete result if hasMore=true. No confirmation is needed for this read. For QUEUED or RUNNING use get_execution.

start_course_student_import

ChatGPT
Start a ready student import/list after inspecting validation results and resolving requested settings. This creates school users and enrollments or updates existing profiles according to mode. Only explicit purchased changes payment status; tariff alone never does. Does not send invitation emails or create payments. Invalid/duplicate rows are reported, never silently claimed imported. Uses the existing background import worker. Tool SUCCEEDED is not import completion: poll get_course_student_import until completed/completed_with_errors/failed and report counts/errors. Do not blindly repeat unknown writes.

unpublish_course

ChatGPT
Only for an explicit request to unpublish/deactivate the course, not to hide it from a listing. Sets course active=false, affecting student access. Does not delete the course, materials, files or students and does not unpublish the landing page. Use expectedRevision from get_course. Preserve requestId and poll get_execution.

unpublish_course_element

ChatGPT
Explicitly unpublish only this material; flag applies to all languages. Preserve draft, published content, forms, files and student records. For a module, children flags remain unchanged. Does not deactivate the course. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

unpublish_page

ChatGPT
Only for an explicit request to unpublish/hide this page. Read get_page first and pass its pageUuid, lang and expectedRevision. Hides the page in ALL languages, even when lang is supplied. Preserves published content, pending drafts, metadata and shared forms; does not delete the page or restore a default template. On CONFLICT reread before proceeding. Preserve requestId and poll get_execution. Republishing later via publish_page publishes any pending draft in the selected language.

update_course

ChatGPT
Read get_course first, then send only requested settings and its expectedRevision. Changes take effect immediately; there is no settings draft. Can switch paid/free and hasPaymentPlans tariff mode without deleting tariff entities/assignments. A paid course without tariffs needs a positive price; tariff prices/access durations use separate tariff commands. Can switch accessType=all/days_after_register/scheduled; scheduled requires a calendar plan first and can affect existing student access. Stored schedules/plans remain intact. Material dates and calendar entities use separate schedule commands. Cannot change publication, material content, tariff entities/assignments, students or design here. On CONFLICT reread and reapply. Preserve requestId; poll get_execution after uncertain outcomes.

update_course_design_config

ChatGPT
Change active course-material styles only for the user-selected scope. Recommend shared parent=courses for consistency; it affects inheriting courses, including future ones, but preserves course overrides. For an explicit only-this-course request use parent=course and courseUuid. If scope is ambiguous explain both and ask before writing; never silently widen a single-course request. Read get_course_design_config for the same target, preserve unrelated generalConfig fields and use its expectedRevision. Active immediately; no draft or publication. Does not change site/landing styles, material content, access or students. For previews use preview_design context=course without saving. QUEUED/RUNNING: poll get_execution; preserve requestId after uncertain outcomes.

update_course_element

ChatGPT
Read get_course_element first; use expectedRevision. Change only requested settings and/or native draft page. Preserve all unrelated fields and form identities. Settings apply immediately; design waits for separate publication. generalConfig is validation only. Course, tariffs, schedule, parent, type and files cannot be replaced through settings. QUEUED/RUNNING: poll get_execution. Preserve requestId on timeout/UNKNOWN; never blindly retry writes.

update_course_group

ChatGPT
Rename a course group or change its schedule using its current revision from get_course_groups. Cannot change schedule of a group containing students. Does not edit student profiles or publish courses.

update_course_material_schedule

ChatGPT
Read get_course_material_schedule with the same mode/plan/language and use its revision. Patch 1-200 materials; omitted boundaries remain unchanged. days uses dayStart/dayEnd (0/null clears); calendar uses dateStart/dateEnd (null clears) with ISO 8601 offsets. Course accessType must match; switch explicitly through update_course first and reread. includeDescendants=true applies recursively to selected modules. Effective immediately; other plans, materials, files, tariffs and students remain unchanged. On CONFLICT reread. Preserve requestId and poll get_execution.

update_course_registration_fields

ChatGPT
First read get_course_registration_fields. Send the complete desired ordered list of existing CRM field IDs and isRequired flags; preserve unrelated entries. [] clears only this course list. IDs must belong to the connected school. Does not create/delete CRM fields, alter saved values or change publication/registration mode. Changes apply immediately. Preserve requestId and poll get_execution on uncertain outcomes.

update_course_schedule

ChatGPT
Read get_course_schedules and use the selected plan item revision. Send only requested settings with ISO 8601 dates and timezone offsets. Omitted dates/flags remain unchanged, including registration dates when education dates change. Effective immediately; no student enrollment/reassignment, material dates, publication or other plans are changed. On CONFLICT reread. Preserve requestId and poll get_execution.

update_course_students

ChatGPT
Update 1-100 existing StudentProfiles in one course. Read get_course_students for UUIDs/revisions first. Only supplied settings change. Tariff never changes isPaid. Payment flag requires explicit user instruction; no payment is collected. Never change user name/email/password. Incompatible group/schedule is rejected; explicitly change/clear group when moving schedules. All targets validated before writing. Does not enroll new students.

update_course_tariff

ChatGPT
Read get_course_tariffs first and use the selected tariff item revision. Send only requested settings. Price/access changes apply immediately, not as a draft. Other translations, material assignments, students, schedules and publication remain unchanged. Hide via isHidden, no deletion command. On CONFLICT reread; preserve requestId and poll get_execution.

update_course_tariff_materials

ChatGPT
Read get_course_tariff_materials first. Add/remove one tariff for 1-200 selected material UUIDs with included=true/false. Other tariff links stay unchanged. includeDescendants=true applies recursively to selected modules, including nested modules; omitted/false targets exact UUIDs. Parent modules are not added automatically. Effective immediately; no content, publication, student, schedule or file changes. On CONFLICT reread; preserve requestId and poll get_execution.

update_course_test_questions

ChatGPT
Read get_course_test first. Apply 1–100 explicit operations atomically: add_question from a verified bank, add_random bank/count, update item points/count, move or remove an item. Settings apply immediately. Remove only detaches from this test, never deletes bank questions or attempts. Points belong to this test, independently of bank points. Fixed duplicates fail; random selections follow native bank sampling and may overlap fixed selections. Do not silently replace existing items. QUEUED/RUNNING: poll get_execution. Preserve requestId after timeouts or UNKNOWN; never blindly repeat writes.

update_custom_field

ChatGPT
Read get_custom_fields and use the selected item revision. Send only requested name, placeholder, icon, visibility or select options; existing type cannot change. Definition changes apply wherever this field is used, including other courses. Preserve option values when renaming labels. Does not delete fields or change saved contact values or course required flags. Poll get_execution after timeouts.

update_page

ChatGPT
Read get_page first. Replace only the requested native draft design using its expectedRevision; preserve all untouched fields and existing form/question identities from this same page. New forms/questions omit entity IDs. Omitted page/metadata fields remain unchanged. Metadata (title/description/alias/img) takes effect immediately; design stays in draft and never publishes. generalConfig is validation context only. On CONFLICT reread and reapply intended edits. Preserve requestId and poll get_execution; no blind retries.

update_question

ChatGPT
Edit one verified question after get_question, using its revision and existing answer UUIDs. Supplied content fully replaces that content, using supported formatting; omitted fields remain unchanged. Add answers by omitting UUID; remove only through explicit removal UUID arrays. Omission from answers never deletes. Do not change the question type. Student-used questions cannot change scoring or answer composition. Show the target and proposed changes if the user's request is ambiguous. Stale revision requires rereading and reconciling; never blindly overwrite. No uploads or bulk edits.

update_site_config

ChatGPT
Only when the user asks to change the overall school site styles. Read get_site_config first; supply its expectedRevision and the complete updated generalConfig, preserving unrelated fields. This updates the ACTIVE site-wide design system immediately, with no draft, and can affect every page. Do not use merely to preview a page-specific proposal. On CONFLICT reread. Preserve requestId and poll get_execution.

update_site_section

ChatGPT
Read get_site_sections first; replace the complete localized header or native footer using its exact revision. Header supports main menu labels, order, page/link actions and one submenu level, plus localized CTA. Footer requires generalConfig from get_site_config for validation only. Preserve unrelated fields and linked form/question identities; new forms omit entity IDs. Saves AND publishes only this section/language atomically. There is no draft/publication mode. Does not edit/publish pages or change shared logo/styles. On CONFLICT reread. Preserve requestId and poll get_execution; never retry an uncertain write.

validate_design

ChatGPT
Check native {page, generalConfig} without saving, changing or normalizing it. Checks core constructor schemas, identities/layout links and actual fine-editor schemas for the host profiles, plus known jsonReader limitations. Returns valid, errors/warnings with JSON Pointer paths and catalog versions. This is not visual QA, HTML/CSS sanitization or verification that external assets/links work. Fix errors before preview; inspect warnings. For course materials pass context=course with the current course generalConfig; task reply forms use this context. General viewer does not verify student access to private video. Maximum design JSON is 480 KiB.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/westudy"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/westudy)