Author courses from any MCP client

scope read:coursesread-only

List the organization's courses (id, title, status, editor_mode, dates), newest first, one page at a time. Returns {courses, total, nextCursor}: total counts every match; while nextCursor is not null, call again with cursor: nextCursor to get the rest — never assume one page is the whole library. Filter by title with query.

Input

{
  "type": "object",
  "properties": {
    "status": {
      "type": "string",
      "enum": [
        "draft",
        "published",
        "archived"
      ],
      "description": "Optional status filter"
    },
    "query": {
      "type": "string",
      "description": "Optional: only courses whose title contains this text (case-insensitive)"
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 100,
      "description": "Page size (default 50, max 100)"
    },
    "cursor": {
      "type": "string",
      "description": "nextCursor from the previous page; omit for the first page"
    }
  },
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "list_courses",
    "arguments": {
      "status": "draft",
      "query": "forklift",
      "limit": 1
    }
  }
}

Example result (structuredContent)

{
  "courses": [
    {
      "id": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Forklift Safety Refresher",
      "description": "Annual certification refresher for warehouse staff.",
      "status": "draft",
      "editor_mode": "classic",
      "created_at": "2026-09-20T09:00:00.000Z",
      "updated_at": "2026-09-25T10:00:00.000Z"
    }
  ],
  "total": 3,
  "nextCursor": "eyJvIjoxfQ"
}

Side effects

None.

Errors

• cursor is not one list_courses returned; omit it to start from the first page.
• status must be one of <statuses>.

Notes

• Pages hold 50 courses by default (limit up to 100), newest created first; the order does not change when a course is edited, so paging never skips or repeats one.
• total counts every course matching the filters. Keep calling with cursor: nextCursor until nextCursor is null.

scope read:coursesread-only

Get a course's full outline: modules, lessons, and each lesson's block list (ids and types, not content). Use get_lesson to read block content. A Scenes course (kind: "scenes" — imported Storyline/Captivate courses, slide-based) returns its slides instead: id, name, order, objectCount, layerNames and a notes excerpt; read a slide's objects with get_scene_slide.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_course",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11"
    }
  }
}

Example result (structuredContent)

{
  "id": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "title": "Forklift Safety Refresher",
  "description": "Annual certification refresher for warehouse staff.",
  "status": "draft",
  "passing_score": 80,
  "language": "en",
  "flow_mode": "linear",
  "updated_at": "2026-09-25T10:00:00.000Z",
  "modules": [
    {
      "id": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
      "title": "Pre-operation checks",
      "sort_order": 0,
      "lessons": [
        {
          "id": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
          "title": "The daily inspection",
          "sort_order": 0,
          "blocks": [
            {
              "id": "f7410216-805d-4e9f-b142-cd3e6f708166",
              "type": "heading",
              "sort_order": 0
            },
            {
              "id": "0852132a-916e-4fa0-8253-de4f70819277",
              "type": "image",
              "sort_order": 1
            },
            {
              "id": "19632438-a27f-4ab1-9364-ef508192a388",
              "type": "multiple_choice",
              "sort_order": 2
            }
          ]
        }
      ]
    }
  ],
  "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "previewUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11/preview"
}

Side effects

None.

Errors

• Course <courseId> not found in your organization.

Notes

• Blocks are listed by id and type only; read their content with get_lesson.

scope write:courseswrites

Create a course, optionally with its whole structure in one call: modules containing lessons containing blocks. This is the best way to build a course — draft the full structure first, then create it in one shot and refine with add_block/update_block. Read the tutorials (resources) before your first course: they cover which blocks to use and how to write good content for each. Courses start as drafts; publish with update_course.

Input

{
  "type": "object",
  "properties": {
    "title": {
      "type": "string",
      "description": "Course title (1–255 chars, no control characters)"
    },
    "description": {
      "type": "string",
      "description": "Optional course description shown in the catalog"
    },
    "structure": {
      "type": "object",
      "description": "Optional full course structure, created with the course in one transaction (all or nothing)",
      "properties": {
        "modules": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "title": {
                "type": "string",
                "description": "Module title (1–255 chars)"
              },
              "lessons": {
                "type": "array",
                "items": {
                  "type": "object",
                  "properties": {
                    "title": {
                      "type": "string",
                      "description": "Lesson title (1–255 chars)"
                    },
                    "questionPoolSize": {
                      "type": [
                        "integer",
                        "null"
                      ],
                      "minimum": 1,
                      "maximum": 1000,
                      "description": "Question pool: how many of the lesson's question blocks each learner attempt gets, drawn at random (stable when the attempt is resumed; scored out of the drawn questions only). Omit or null for every question."
                    },
                    "blocks": {
                      "type": "array",
                      "items": {
                        "type": "object",
                        "description": "One content block. Call list_block_types for the content schema of each type.",
                        "properties": {
                          "type": {
                            "type": "string",
                            "description": "Block type, e.g. \"paragraph\", \"heading\", \"multiple_choice\""
                          },
                          "content": {
                            "type": "object",
                            "description": "Block content matching that type's schema",
                            "additionalProperties": true
                          },
                          "settings": {
                            "type": "object",
                            "description": "Optional presentation settings",
                            "additionalProperties": true
                          }
                        },
                        "required": [
                          "type"
                        ],
                        "additionalProperties": false
                      }
                    }
                  },
                  "required": [
                    "title"
                  ],
                  "additionalProperties": false
                }
              }
            },
            "required": [
              "title"
            ],
            "additionalProperties": false
          }
        }
      },
      "additionalProperties": false
    }
  },
  "required": [
    "title"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "create_course",
    "arguments": {
      "title": "Forklift Safety Refresher",
      "description": "Annual certification refresher for warehouse staff.",
      "structure": {
        "modules": [
          {
            "title": "Pre-operation checks",
            "lessons": [
              {
                "title": "The daily inspection",
                "blocks": [
                  {
                    "type": "heading",
                    "content": {
                      "text": "Module 1: Hazard Identification",
                      "level": "h2"
                    }
                  },
                  {
                    "type": "image",
                    "content": {
                      "url": "https://example.com/assets/forklift-inspection.jpg",
                      "alt": "A warehouse worker checking a forklift's tires during a pre-shift inspection",
                      "caption": "Daily pre-shift forklift inspection"
                    }
                  },
                  {
                    "type": "multiple_choice",
                    "content": {
                      "question": "<p>You spot a small chemical spill in an aisle. What should you do first?</p>",
                      "options": [
                        {
                          "text": "Mop it up yourself right away",
                          "correct": false,
                          "feedback": "Not until you know what the chemical is — check the label and SDS first."
                        },
                        {
                          "text": "Secure the area and check the safety data sheet",
                          "correct": true,
                          "feedback": "Right — identify the substance before anyone touches it."
                        },
                        {
                          "text": "Ignore it if it is smaller than your hand",
                          "correct": false,
                          "feedback": "Size doesn't determine hazard — some chemicals are dangerous in tiny amounts."
                        }
                      ]
                    }
                  }
                ]
              }
            ]
          }
        ]
      }
    }
  }
}

Example result (structuredContent)

{
  "id": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "title": "Forklift Safety Refresher",
  "description": "Annual certification refresher for warehouse staff.",
  "status": "draft",
  "passing_score": 80,
  "language": "en",
  "flow_mode": "linear",
  "updated_at": "2026-09-25T10:00:00.000Z",
  "modules": [
    {
      "id": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
      "title": "Pre-operation checks",
      "sort_order": 0,
      "lessons": [
        {
          "id": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
          "title": "The daily inspection",
          "sort_order": 0,
          "blocks": [
            {
              "id": "f7410216-805d-4e9f-b142-cd3e6f708166",
              "type": "heading",
              "sort_order": 0
            },
            {
              "id": "0852132a-916e-4fa0-8253-de4f70819277",
              "type": "image",
              "sort_order": 1
            },
            {
              "id": "19632438-a27f-4ab1-9364-ef508192a388",
              "type": "multiple_choice",
              "sort_order": 2
            }
          ]
        }
      ]
    }
  ],
  "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "previewUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11/preview"
}

Side effects

Creates a draft course with its modules, lessons and blocks. A course with no modules gets "Module 1" / "Lesson 1"; a module with no lessons gets "Lesson 1".

Errors

• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• description must be 5000 characters or fewer
• structure.modules must be an array
• structure.modules[<i>].lessons[<j>].blocks[<k>] must be an object with a string type
• Unknown block type "<type>". Valid types: <list>. Call list_block_types for each type's content schema.
• "<lesson title>": <type> needs content.<field> (<type>): <field description> Call list_block_types for the full reference.
• "<lesson title>": content.<field> must be a <type>, not <actual>. Call list_block_types for the full reference.
• "<lesson title>": multiple_choice needs exactly one option with correct: true (found <n>). Use multi_select for several.
• block content too large (max 512KB)
• block settings too large (max 512KB)

Notes

• Titles (course, module, lesson) are trimmed and must be 1–255 characters with no control characters. Errors name the field, e.g. structure.modules[0].lessons[2].title.
• Every title and block is validated before anything is written, and the course is then created in one database transaction: an error result means nothing was created, so it is safe to fix the call and retry.
• Soft problems (unknown content fields, values outside an enum) do not fail the call; they come back in a warnings array on the result.

scope write:courseswritesdestructive

Update a course's title, description, passing score (0–100), or status. Setting status to "published" makes it available to learners.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "New title (1–255 chars)"
    },
    "description": {
      "type": "string",
      "description": "New description"
    },
    "status": {
      "type": "string",
      "enum": [
        "draft",
        "published",
        "archived"
      ]
    },
    "passing_score": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100,
      "description": "Quiz pass threshold in percent"
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_course",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "status": "published",
      "passing_score": 80
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "applied": {
    "status": "published",
    "passing_score": 80
  }
}

Side effects

Updates the course row and pushes the change to any editor that has the course open. status "published" makes the course available to learners and fires the organization's `course.published` webhook, exactly as publishing in the editor does.

Errors

• Course <courseId> not found in your organization.
• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• description must be 5000 characters or fewer
• status must be "draft", "published", or "archived"
• passing_score must be an integer 0–100
• Nothing to update.

scope write:courseswrites

Append a module (a chapter grouping lessons) to a course.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "Module title (1–255 chars)"
    }
  },
  "required": [
    "courseId",
    "title"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "add_module",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Operating safely"
    }
  }
}

Example result (structuredContent)

{
  "id": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
  "title": "Operating safely",
  "sort_order": 1
}

Side effects

Appends a module at the end of the course.

Errors

• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• Course <courseId> not found in your organization.

Notes

• Reorder with update_module; the module starts with no lessons.

scope write:courseswrites

Append a lesson to a module. A lesson is a page of blocks the learner reads top to bottom.

Input

{
  "type": "object",
  "properties": {
    "moduleId": {
      "type": "string",
      "description": "Module id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "Lesson title (1–255 chars)"
    }
  },
  "required": [
    "moduleId",
    "title"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "add_lesson",
    "arguments": {
      "moduleId": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
      "title": "Check your knowledge"
    }
  }
}

Example result (structuredContent)

{
  "id": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
  "title": "Check your knowledge",
  "module_id": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
  "sort_order": 1
}

Side effects

Appends an empty lesson at the end of the module.

Errors

• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• Module <moduleId> not found in your organization.

scope read:coursesread-only

Get a lesson with the full content and settings of every block, in order.

Input

{
  "type": "object",
  "properties": {
    "lessonId": {
      "type": "string",
      "description": "Lesson id (uuid)"
    }
  },
  "required": [
    "lessonId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_lesson",
    "arguments": {
      "lessonId": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33"
    }
  }
}

Example result (structuredContent)

{
  "id": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
  "title": "The daily inspection",
  "module_id": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
  "sort_order": 0,
  "blocks": [
    {
      "id": "f7410216-805d-4e9f-b142-cd3e6f708166",
      "type": "heading",
      "content": {
        "text": "Module 1: Hazard Identification",
        "level": "h2"
      },
      "settings": {},
      "sort_order": 0
    },
    {
      "id": "0852132a-916e-4fa0-8253-de4f70819277",
      "type": "image",
      "content": {
        "url": "https://example.com/assets/forklift-inspection.jpg",
        "alt": "A warehouse worker checking a forklift's tires during a pre-shift inspection",
        "caption": "Daily pre-shift forklift inspection"
      },
      "settings": {},
      "sort_order": 1
    }
  ]
}

Side effects

None.

Errors

• Lesson <lessonId> not found in your organization.

scope write:courseswrites

Append one content block to a lesson (or insert at a position). The content object must match the block type's schema — call list_block_types first if unsure.

Input

{
  "type": "object",
  "properties": {
    "lessonId": {
      "type": "string",
      "description": "Lesson id (uuid)"
    },
    "type": {
      "type": "string",
      "description": "Block type"
    },
    "content": {
      "type": "object",
      "description": "Content matching the block type's schema",
      "additionalProperties": true
    },
    "settings": {
      "type": "object",
      "description": "Optional presentation settings",
      "additionalProperties": true
    },
    "position": {
      "type": "integer",
      "minimum": 0,
      "description": "Optional sort position; defaults to the end"
    }
  },
  "required": [
    "lessonId",
    "type",
    "content"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "add_block",
    "arguments": {
      "lessonId": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
      "type": "multiple_choice",
      "content": {
        "question": "<p>You spot a small chemical spill in an aisle. What should you do first?</p>",
        "options": [
          {
            "text": "Mop it up yourself right away",
            "correct": false,
            "feedback": "Not until you know what the chemical is — check the label and SDS first."
          },
          {
            "text": "Secure the area and check the safety data sheet",
            "correct": true,
            "feedback": "Right — identify the substance before anyone touches it."
          },
          {
            "text": "Ignore it if it is smaller than your hand",
            "correct": false,
            "feedback": "Size doesn't determine hazard — some chemicals are dangerous in tiny amounts."
          }
        ]
      },
      "position": 0
    }
  }
}

Example result (structuredContent)

{
  "id": "19632438-a27f-4ab1-9364-ef508192a388",
  "type": "multiple_choice",
  "sort_order": 0
}

Side effects

Inserts the block (at the end, or at position with the siblings renumbered 0..n-1) and pushes it to any editor that has the lesson open.

Errors

• Unknown block type "<type>". Valid types: <list>. Call list_block_types for each type's content schema.
• <type> needs content.<field> (<type>): <field description> Call list_block_types for the full reference.
• content.<field> must be a <type>, not <actual>. Call list_block_types for the full reference.
• multiple_choice needs exactly one option with correct: true (found <n>). Use multi_select for several.
• content too large (max 512KB)
• settings too large (max 512KB)
• Lesson <lessonId> not found in your organization.
• position must be a non-negative integer

Notes

• Soft content problems come back as a warnings array on the result; the block is still created.
• Every argument, position included, is checked before anything is written: an error result means no block was added.

scope write:courseswritesdestructive

Replace a block's content and/or settings, or move it (sort_order). Content is replaced whole, not merged — send the complete object.

Input

{
  "type": "object",
  "properties": {
    "blockId": {
      "type": "string",
      "description": "Block id (uuid)"
    },
    "content": {
      "type": "object",
      "description": "Complete replacement content",
      "additionalProperties": true
    },
    "settings": {
      "type": "object",
      "description": "Complete replacement settings",
      "additionalProperties": true
    },
    "sort_order": {
      "type": "integer",
      "minimum": 0
    }
  },
  "required": [
    "blockId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_block",
    "arguments": {
      "blockId": "0852132a-916e-4fa0-8253-de4f70819277",
      "content": {
        "url": "https://example.com/assets/forklift-inspection.jpg",
        "alt": "A worker checking a forklift's tyre pressure before a shift",
        "caption": "Daily pre-shift forklift inspection"
      }
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "blockId": "0852132a-916e-4fa0-8253-de4f70819277",
  "applied": [
    "content"
  ]
}

Side effects

Replaces content and/or settings whole (no merge) and/or moves the block, then pushes the change to open editors.

Errors

• <type> needs content.<field> (<type>): <field description> Call list_block_types for the full reference.
• content.<field> must be a <type>, not <actual>. Call list_block_types for the full reference.
• multiple_choice needs exactly one option with correct: true (found <n>). Use multi_select for several.
• Block <blockId> not found in your organization.
• content too large (max 512KB)
• settings too large (max 512KB)
• sort_order must be a non-negative integer (the 0-based position)
• Nothing to update.

Notes

• Read the block with get_lesson, change what you need, and send the complete object back — omitted fields are dropped.

scope write:courseswritesdestructive

Permanently delete one block from a lesson.

Input

{
  "type": "object",
  "properties": {
    "blockId": {
      "type": "string",
      "description": "Block id (uuid)"
    }
  },
  "required": [
    "blockId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "delete_block",
    "arguments": {
      "blockId": "19632438-a27f-4ab1-9364-ef508192a388"
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "deleted": "19632438-a27f-4ab1-9364-ef508192a388"
}

Side effects

Permanently deletes the block. There is no undo over MCP.

Errors

• Block <blockId> not found in your organization.

scope read:coursesread-only

The authoritative reference for all 37 block types: category, description, exact content field schema, and a realistic example for each. Pass a type to get just that block, or a category (Text, Media, Layout, Interactive, Assessment) to browse.

Input

{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "description": "Optional: one block type to describe in full"
    },
    "category": {
      "type": "string",
      "enum": [
        "Text",
        "Media",
        "Layout",
        "Interactive",
        "Assessment"
      ]
    }
  },
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "list_block_types",
    "arguments": {
      "type": "multiple_choice"
    }
  }
}

Example result (structuredContent)

{
  "type": "multiple_choice",
  "label": "Multiple Choice",
  "category": "Assessment",
  "description": "Single correct answer quiz",
  "contentFields": [
    {
      "name": "question",
      "type": "string",
      "required": true,
      "description": "Question text as an HTML string (TipTap)."
    },
    {
      "name": "options",
      "type": "{ text: string; correct: boolean; feedback: string }[]",
      "required": true,
      "description": "Answer options. Exactly ONE option should have correct: true (the editor enforces this with radio buttons; the grader accepts whichever index has correct === true). feedback is per-option text shown to the learner after submitting when settings.showFeedback is on; use an empty string when there is none. The editor caps options at 6. A true/false question is a multiple_choice with the two options \"True\" and \"False\"; there is no separate type."
    }
  ],
  "example": {
    "question": "<p>You spot a small chemical spill in an aisle. What should you do first?</p>",
    "options": [
      {
        "text": "Mop it up yourself right away",
        "correct": false,
        "feedback": "Not until you know what the chemical is — check the label and SDS first."
      },
      {
        "text": "Secure the area and check the safety data sheet",
        "correct": true,
        "feedback": "Right — identify the substance before anyone touches it."
      },
      {
        "text": "Ignore it if it is smaller than your hand",
        "correct": false,
        "feedback": "Size doesn't determine hazard — some chemicals are dangerous in tiny amounts."
      }
    ]
  },
  "notes": "GRADER AUTHORITY (lib/grading.ts): the learner response is the chosen option INDEX (number); options[chosen].correct === true wins. correct and feedback are stripped from learner-facing content by redactAnswers, so grading only happens server-side. settings.shuffleOptions: boolean shuffles display order; settings.showFeedback: boolean (default true) shows per-option feedback."
}

Side effects

None.

Errors

• Unknown type "<type>". Valid types: <list>.

Notes

• Without arguments (or with category) it returns a compact list: type, category, description and contentFields as "name?: type" strings.

scope write:courseswritesdestructive

Rename a module or change its position (sort_order) within the course.

Input

{
  "type": "object",
  "properties": {
    "moduleId": {
      "type": "string",
      "description": "Module id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "New title (1–255 chars)"
    },
    "sort_order": {
      "type": "integer",
      "minimum": 0,
      "description": "New position among the course's modules (0-based)"
    }
  },
  "required": [
    "moduleId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_module",
    "arguments": {
      "moduleId": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
      "title": "Before you drive",
      "sort_order": 0
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "moduleId": "8b2d4e61-1c3f-4a7e-b5d9-0e6f2a7c3d22",
  "applied": {
    "title": "Before you drive",
    "sort_order": 0
  }
}

Side effects

Renames and/or moves the module; sibling modules are renumbered 0..n-1.

Errors

• Module <moduleId> not found in your organization.
• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• sort_order must be a non-negative integer (the 0-based position)
• Nothing to update.

scope write:courseswritesdestructive

Rename a lesson, change its position (sort_order) within its module, or make it a question pool (question_pool_size: each learner attempt gets that many of the lesson's question blocks at random, is scored out of those, and keeps the same draw when resumed; null gives every question).

Input

{
  "type": "object",
  "properties": {
    "lessonId": {
      "type": "string",
      "description": "Lesson id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "New title (1–255 chars)"
    },
    "sort_order": {
      "type": "integer",
      "minimum": 0,
      "description": "New position among the module's lessons (0-based)"
    },
    "question_pool_size": {
      "type": [
        "integer",
        "null"
      ],
      "minimum": 1,
      "maximum": 1000,
      "description": "Question pool: how many of the lesson's question blocks each learner attempt gets, drawn at random (stable when the attempt is resumed; scored out of the drawn questions only). Omit or null for every question."
    }
  },
  "required": [
    "lessonId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_lesson",
    "arguments": {
      "lessonId": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
      "sort_order": 1
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "lessonId": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
  "applied": {
    "sort_order": 1
  }
}

Side effects

Renames and/or moves the lesson within its module; siblings are renumbered 0..n-1.

Errors

• Lesson <lessonId> not found in your organization.
• <field> must be a string
• <field> must be 1–255 characters
• <field> must not contain control characters (tabs, newlines…)
• sort_order must be a non-negative integer (the 0-based position)
• Nothing to update.

scope write:courseswritesdestructive

Turn a course into an adaptive path (or back to linear). Describe the graph in lessons and conditions — nodes reference lesson ids from get_course, edges carry conditions (assessment_score on the lesson the edge leaves, or none for the one fallback edge). Layout, geometry, and editor metadata are computed for you. Read opencourse://tutorials/branching first; pass flow_mode "linear" with no graph to switch branching off.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "flow_mode": {
      "type": "string",
      "enum": [
        "linear",
        "branching"
      ]
    },
    "graph": {
      "type": "object",
      "description": "Required when flow_mode is \"branching\"",
      "properties": {
        "nodes": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "lessonId": {
                "type": "string",
                "description": "A lesson id from get_course"
              },
              "isEntry": {
                "type": "boolean",
                "description": "Exactly one node must be the entry"
              },
              "isExit": {
                "type": "boolean",
                "description": "At least one node must be an exit"
              }
            },
            "required": [
              "lessonId"
            ],
            "additionalProperties": false
          }
        },
        "edges": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "from": {
                "type": "string",
                "description": "Source lesson id"
              },
              "to": {
                "type": "string",
                "description": "Target lesson id"
              },
              "condition": {
                "type": "object",
                "description": "Condition object, or omit for an unconditional edge. See the branching tutorial.",
                "additionalProperties": true
              },
              "label": {
                "type": "string",
                "description": "Short label shown on the edge, e.g. \"score ≥ 80%\""
              },
              "priority": {
                "type": "integer",
                "description": "Lower evaluates first when several edges leave one node"
              }
            },
            "required": [
              "from",
              "to"
            ],
            "additionalProperties": false
          }
        }
      },
      "required": [
        "nodes",
        "edges"
      ],
      "additionalProperties": false
    }
  },
  "required": [
    "courseId",
    "flow_mode"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "set_flow",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "flow_mode": "branching",
      "graph": {
        "nodes": [
          {
            "lessonId": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
            "isEntry": true
          },
          {
            "lessonId": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
            "isExit": true
          },
          {
            "lessonId": "e6309105-7f4c-4d8e-a031-bc2d5e6f7055"
          }
        ],
        "edges": [
          {
            "from": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
            "to": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44"
          },
          {
            "from": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
            "to": "e6309105-7f4c-4d8e-a031-bc2d5e6f7055",
            "condition": {
              "type": "assessment_score",
              "operator": "<",
              "value": 80
            },
            "label": "score < 80%"
          },
          {
            "from": "e6309105-7f4c-4d8e-a031-bc2d5e6f7055",
            "to": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
            "label": "try again"
          }
        ]
      }
    }
  }
}

Example result (structuredContent)

{
  "ok": true,
  "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "flow_mode": "branching",
  "nodes": 3,
  "edges": 3,
  "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11"
}

Side effects

Saves the course's flow mode and (for branching) its graph, laid out automatically, and pushes it to open editors. Switching to linear keeps the stored graph so it can be switched back on.

Errors

• Course <courseId> not found in your organization.
• flow_mode must be "linear" or "branching"
• flow_mode "branching" requires a graph
• The graph needs at least one node.
• Each lesson can appear in the graph only once.
• Node lesson <lessonId> is not a lesson of this course. Call get_course for valid lesson ids.
• Exactly one node must have isEntry: true (got <n>). The entry is where every learner starts.
• At least one node must have isExit: true — the lesson(s) where the course can complete.
• edges[<i>].from (<id>) is not a node in the graph
• edges[<i>].to (<id>) is not a node in the graph
• edges[<i>].condition: compound conditions are not supported; use one assessment_score condition per edge.
• edges[<i>].condition: condition type "<type>" is not supported. Valid: <always, assessment_score>.
• edges[<i>].condition: assessment_score needs an operator (<operators>)
• edges[<i>].condition: assessment_score needs a numeric value (0–100)
• edges[<i>].condition: assessment_score tests the lesson this edge leaves, so its lessonId must be <from> (or omitted).
• "<lesson>" has <n> unconditional edges; a lesson can have only one fallback.
• "<lesson>" is an exit but has an unconditional edge out of it, so the course could never end there. Remove that edge, or make it conditional.
• "<lesson>" is not an exit and has no unconditional edge, so a learner whose score matches no condition would be stranded there. Add a fallback edge, or mark it isExit.
• Unreachable node(s): <titles>. Every node needs a path from the entry.

Notes

• See “Branching flows” below for the portability rules these errors enforce.

scope read:coursesread-only

Check a whole course for accessibility and design problems: useless or missing alt text, uncaptioned video, broken quiz answer keys, heading-order violations, empty or overlong lessons. Returns findings by severity (error, warning, suggestion; Scenes courses report error, warning, info). Run this after building and fix every error before telling the human you're done. On a Scenes course it runs the same preflight the Scenes export runs (unnamed controls, images without alt text, broken trigger references, contrast, quiz keys…): findings carry slideId/objectId, and byRule says which tool fixes each rule (update_scene_object, update_scene_slide, or the editor). Findings are capped at 150 per call; filter with slideId or severity.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "slideId": {
      "type": "string",
      "description": "Scenes courses only: just this slide's findings"
    },
    "severity": {
      "type": "string",
      "enum": [
        "error",
        "warning",
        "info"
      ],
      "description": "Scenes courses only: just this severity"
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "audit_course",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11"
    }
  }
}

Example result (structuredContent)

{
  "findings": [
    {
      "severity": "suggestion",
      "rule": "course-description",
      "message": "The course has no description. Learners see it in the catalog — one or two sentences on what they'll be able to do."
    },
    {
      "severity": "warning",
      "rule": "heading-order",
      "message": "Heading level jumps to h3 after h1. Levels must descend without skipping.",
      "lessonId": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
      "lessonTitle": "The daily inspection",
      "blockId": "f7410216-805d-4e9f-b142-cd3e6f708166",
      "blockType": "heading"
    },
    {
      "severity": "error",
      "rule": "alt-text",
      "message": "Alt text \"IMG_0042.jpg\" is a file name or a placeholder, not a description. Write what the image teaches.",
      "lessonId": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
      "lessonTitle": "The daily inspection",
      "blockId": "0852132a-916e-4fa0-8253-de4f70819277",
      "blockType": "image"
    },
    {
      "severity": "warning",
      "rule": "empty-lesson",
      "message": "Lesson has no blocks.",
      "lessonId": "d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44",
      "lessonTitle": "Check your knowledge"
    }
  ],
  "summary": {
    "error": 1,
    "warning": 2,
    "suggestion": 1
  }
}

Side effects

None.

Errors

• Course <courseId> not found in your organization.

Notes

• Rules: course-description, empty-lesson, lesson-length, block-content, alt-text (skipped for an image with decorative: true), alt-text-reused, video-captions, video-alternative, audio-transcript, heading-order, quiz-options, quiz-answer-key, hotspot-area-name, custom-interaction-title.
• Severities are error, warning and suggestion; summary counts each.
• On a Scenes course the result is the Scenes export preflight instead (kind: "scenes"): severities error, warning and info; findings with slideId/objectId; byRule with each rule's message, count and fixWith (update_scene_object, update_scene_slide or editor). At most 150 findings per call; slideId and severity filter.

scope read:coursesread-only

Build a standards-compliant package (SCORM 1.2, SCORM 2004, cmi5, or xAPI) and return a download link that works for 24 hours. Hand the link to the human; the package uploads to any LMS. Block and Scenes courses both export. A Scenes course is refused, with the findings, while the preflight finds errors (as the editor's export is): fix them (audit_course says how) or pass force: true only when the human wants a work-in-progress package. When the human will upload the package over a course already in their LMS that this one was imported from ("replace the original", "update the old course", "keep learners' completions"), pass replaceOriginal: true in the original's format; otherwise leave it off, because the default package is a new course to the LMS.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "format": {
      "type": "string",
      "enum": [
        "scorm12",
        "scorm2004",
        "cmi5",
        "xapi"
      ],
      "description": "Defaults to scorm2004"
    },
    "locale": {
      "type": "string",
      "description": "Optional BCP-47 locale to export a stored translation, e.g. \"de\" (block courses only)"
    },
    "languages": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "minItems": 1,
      "description": "Block courses only: one package carrying the course's own language and its translations, with a language picker for the learner (LMS preference, then browser, then the course's language) and a switch in the player. Progress and score do not depend on the language. [\"all\"] takes every complete translation; a list of locales (e.g. [\"fr\", \"de\"]) takes those, complete or not (missing strings show in the course's language). The result's languages.notes name what was left out or incomplete; pass them on. Overrides locale."
    },
    "force": {
      "type": "boolean",
      "description": "Scenes courses only: export even though the preflight finds errors. The result then carries a warning to pass on."
    },
    "replaceOriginal": {
      "type": "boolean",
      "description": "Only for a course imported from an LMS package: keep the original package's identifiers so the LMS can take it as a new version of the old course. Refused when the course was not imported from a package, or when the original had no identity in this format (e.g. cmi5 for a SCORM original): export in the format the refusal names. The result's replaceOriginal.notes say what to check (a changed SCORM version, a multi-part original); pass them on. Tell the human to upload it through the LMS's replace workflow and to try a test copy first."
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "export_course",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "format": "scorm12"
    }
  }
}

Example result (structuredContent)

{
  "downloadUrl": "<signed storage URL>",
  "filename": "Forklift_Safety_Refresher_SCORM1.2.zip",
  "format": "scorm12",
  "sizeBytes": 482113,
  "expiresIn": "24 hours"
}

Side effects

Builds the package and stores the zip in the organization's file storage (under exports/). Annotated read-only because course content is not changed.

Errors

• Course <courseId> not found in your organization.
• This course was not imported from an LMS package, so there is no original identity to keep.
• The original was published as <original format>, not cmi5, so there is no cmi5 identity to keep. Export as <original format> to replace it.

Notes

• format defaults to scorm2004. A locale that is not a valid BCP-47 tag is ignored and the source language is exported.
• Builds can take a while for large courses; the endpoint allows up to 60 seconds per request.
• A Scenes course is built by the same builder as the Scenes editor's export, and refused like it while its preflight finds errors; force: true exports anyway and adds a warning to the result.
• replaceOriginal: true is the editor's "Replace the original in your LMS": for a course imported from an LMS package, the package keeps the original's identifiers (SCORM manifest, organization, item and resource of its first SCO; cmi5 course and first AU id; xAPI activity id) so the LMS can take it as a new version of the old course. Use it only when the human will upload it over that course, in the original's format.
• With replaceOriginal the result carries replaceOriginal: { notes, howTo }: notes flag a changed SCORM version or a multi-part original, howTo is the upload guidance the editor shows. Replacement is verified in Moodle only; tell the human to try a test copy first.

scope read:coursesread-only

Read one slide of a Scenes course: its name and notes, its layers, and per object the id, type, name, layer, hidden flag, words (text: [{ path, label, value }], long values cut at 280 characters), alt_text, aria_label, aria_role, the accessibleName a screen reader will announce (null on a control that has none), and how many states and triggers it has. Geometry and trigger bodies are left out. 100 objects per call; the reply says how to page.

Input

{
  "type": "object",
  "properties": {
    "slideId": {
      "type": "string",
      "description": "Slide id (uuid), from get_course or an audit_course finding"
    },
    "offset": {
      "type": "integer",
      "minimum": 0,
      "description": "Skip this many objects (for slides with more than 100)"
    }
  },
  "required": [
    "slideId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_scene_slide",
    "arguments": {
      "slideId": "3b84465a-c491-4cd3-b586-0172a314c5aa"
    }
  }
}

Example result (structuredContent)

{
  "slide": {
    "id": "3b84465a-c491-4cd3-b586-0172a314c5aa",
    "courseId": "2a743549-b380-4bc2-a475-f0619203b499",
    "name": "Leadership styles",
    "order": 4,
    "notes": ""
  },
  "layers": [
    {
      "id": "<layer id>",
      "name": "Autocratic"
    }
  ],
  "objectCount": 2,
  "objects": [
    {
      "id": "4c95576b-d5a2-4de4-8697-1283b425d6bb",
      "type": "shape",
      "name": "Rectangle 12",
      "interactive": true,
      "accessibleName": null,
      "states": 0,
      "triggers": 1
    },
    {
      "id": "<object id>",
      "type": "text",
      "name": "Text_Caption_95",
      "text": [
        {
          "path": "content.html",
          "label": "Text",
          "value": "Click each style to learn more"
        }
      ],
      "states": 0,
      "triggers": 0
    }
  ]
}

Side effects

None.

Errors

• Slide <slideId> not found in your organization.

Notes

• accessibleName null marks a control a screen reader cannot name: give it an aria_label with update_scene_object.

scope read:coursesread-only

See the pictures on one Scenes slide (up to 6, undescribed ones first, downscaled) so you can write accurate alt text with update_scene_object — never describe an image you have not seen. Each picture comes with its objectId and current alt text.

Input

{
  "type": "object",
  "properties": {
    "slideId": {
      "type": "string",
      "description": "Slide id (uuid), from get_course or an audit_course finding"
    },
    "objectIds": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "maxItems": 6,
      "description": "Only these image objects (from get_scene_slide or the audit)"
    }
  },
  "required": [
    "slideId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "view_scene_images",
    "arguments": {
      "slideId": "3b84465a-c491-4cd3-b586-0172a314c5aa"
    }
  }
}

Example result (structuredContent)

{
  "slide": {
    "id": "3b84465a-c491-4cd3-b586-0172a314c5aa",
    "name": "PMBOK"
  },
  "images": [
    {
      "objectId": "4c95576b-d5a2-4de4-8697-1283b425d6bb",
      "name": "book1",
      "altText": null,
      "decorative": false,
      "image": "shown"
    }
  ],
  "total": 1,
  "next": "Describe each picture with update_scene_object { objectId, alt_text } …"
}

Side effects

None. The content also carries each picture as an MCP image block (JPEG, at most 768 px on its longer side).

Errors

• Slide <slideId> not found in your organization.
• objectIds must be an array of object ids from get_scene_slide.
• None of those objectIds is an image on this slide.

Notes

• Only pictures stored in your organization's media library are shown; any other image is listed as not viewable.
• Text rendered as an image (common in Captivate captions) should get that text as its alt text.

scope write:courseswritesdestructive

Change what an author would retype on one object of a Scenes course: its words (text: [{ path, value }] with a path from get_scene_slide — plain text is fine, it becomes paragraphs), alt_text ("" marks an image decorative), aria_label (the name a screen reader announces), aria_role, name, or hidden. Triggers, variables, states' behavior and geometry are never touched. Open editors update live, and the first edit in a session saves a restorable version.

Input

{
  "type": "object",
  "properties": {
    "objectId": {
      "type": "string",
      "description": "Object id (uuid), from get_scene_slide or an audit_course finding"
    },
    "text": {
      "type": "array",
      "description": "New words for existing text fields",
      "items": {
        "type": "object",
        "properties": {
          "path": {
            "type": "string",
            "description": "A path from get_scene_slide's text list, e.g. \"content.html\" or \"content.options.0.text\""
          },
          "value": {
            "type": "string",
            "description": "The new text (plain text, or HTML for an html path)"
          }
        },
        "required": [
          "path",
          "value"
        ],
        "additionalProperties": false
      }
    },
    "alt_text": {
      "type": [
        "string",
        "null"
      ],
      "description": "Image description (max 1000); \"\" for decorative, null to clear"
    },
    "aria_label": {
      "type": [
        "string",
        "null"
      ],
      "description": "Accessible name for a control or group (max 1000); null to clear"
    },
    "aria_role": {
      "type": [
        "string",
        "null"
      ],
      "enum": [
        "none",
        "button",
        "link",
        "image",
        "heading",
        "text",
        "checkbox",
        "radio",
        null
      ],
      "description": "ARIA role, as the editor's Accessibility panel offers them; null to clear"
    },
    "name": {
      "type": "string",
      "description": "Object name in the editor's layer list (1–255 chars)"
    },
    "hidden": {
      "type": "boolean",
      "description": "Hidden at slide start (a trigger may still show it)"
    },
    "focusable": {
      "type": [
        "boolean",
        "null"
      ],
      "description": "Whether it is a keyboard control. null (the default) infers it: a control when it has a click/hover trigger or is a native control. false for text or images an import wrongly made interactive; true only for a real control the inference misses."
    }
  },
  "required": [
    "objectId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_scene_object",
    "arguments": {
      "objectId": "4c95576b-d5a2-4de4-8697-1283b425d6bb",
      "aria_label": "Show the Autocratic style"
    }
  }
}

Example result (structuredContent)

{
  "id": "4c95576b-d5a2-4de4-8697-1283b425d6bb",
  "courseId": "2a743549-b380-4bc2-a475-f0619203b499",
  "slideId": "3b84465a-c491-4cd3-b586-0172a314c5aa",
  "changed": [
    "aria_label"
  ],
  "versionSaved": "Saved a restorable version \"Before assistant edits\" first.",
  "accessibleName": "Show the Autocratic style"
}

Side effects

Updates the object and broadcasts the change to open Scenes editors. The first edit to a course in 30 minutes saves a version ("Before assistant edits") the author can restore.

Errors

• Object <objectId> not found in your organization.
• "<path>" is not a text field on this object. Use a path from get_scene_slide's text list: <paths>.
• aria_role must be one of <roles> (or null).
• Nothing to change: pass at least one of text, alt_text, aria_label, aria_role, name, hidden, focusable.
• This picture is a control (it has a click trigger), often an invisible click area: decorative would hide it from keyboard users. Give it an aria_label saying what clicking it does.

Notes

• Triggers, variables, states' behavior and geometry cannot be changed over MCP; findings that need them name the editor.

scope write:courseswritesdestructive

Rename a slide of a Scenes course or rewrite its notes (the narration script / speaker notes). Open editors update live.

Input

{
  "type": "object",
  "properties": {
    "slideId": {
      "type": "string",
      "description": "Slide id (uuid)"
    },
    "name": {
      "type": "string",
      "description": "Slide name (1–255 chars)"
    },
    "notes": {
      "type": "string",
      "description": "Slide notes (max 20,000 chars; \"\" clears them)"
    }
  },
  "required": [
    "slideId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "update_scene_slide",
    "arguments": {
      "slideId": "3b84465a-c491-4cd3-b586-0172a314c5aa",
      "name": "Six leadership styles"
    }
  }
}

Example result (structuredContent)

{
  "id": "3b84465a-c491-4cd3-b586-0172a314c5aa",
  "courseId": "2a743549-b380-4bc2-a475-f0619203b499",
  "changed": [
    "name"
  ]
}

Side effects

Updates the slide and broadcasts the change to open Scenes editors; saves a restorable version first, as update_scene_object does.

Errors

• Slide <slideId> not found in your organization.
• Nothing to change: pass name or notes.

scope write:courseswrites

Fetch a public https image, video, audio, or PDF (max 25MB) into the organization's media library and return its permanent URL — use that URL in image/video/audio block content. Counts against the plan's storage quota.

Input

{
  "type": "object",
  "properties": {
    "url": {
      "type": "string",
      "description": "Public https URL of the file to import"
    },
    "filename": {
      "type": "string",
      "description": "Optional filename hint, used for the extension"
    }
  },
  "required": [
    "url"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "upload_asset",
    "arguments": {
      "url": "https://example.com/assets/forklift-inspection.jpg"
    }
  }
}

Example result (structuredContent)

{
  "url": "<public storage URL>",
  "mimeType": "image/jpeg",
  "sizeBytes": 184220
}

Side effects

Stores the file in the organization's media library. Counts against the plan's storage quota.

Errors

• Could not use that URL: <reason> It must be a public https URL.
• Content type "<type>" is not allowed. Images, video, audio, and PDF only.
• This upload would exceed the organization's storage quota. Free space in the media library or upgrade the plan.

Notes

• Reasons include: the URL resolves to a private or internal address, too many redirects (max 3), the source answered with a non-2xx status, or the file is larger than 25MB.

scope write:courseswrites

Import a legacy e-learning package as a new draft course. A published .zip is recognized by its contents: Captivate (Classic 9–2019 and the all-new 12/13), Storyline, Rise 360, Lectora, iSpring, Articulate Studio (recognized; Flash reported), SCORM 1.2/2004, AICC, cmi5, xAPI, Adapt/Evolve, H5P and IMS Common Cartridge/QTI. Source files: .cptx, .story, .awt, .pptx, .h5p, .imscc and Moodle backups (.mbz). Pass a public https url to the file (max 100MB), or the storagePath of a file already staged in the app's course-imports storage. Returns the new courseId and the import report — what could not be recreated — so you can plan fixes: then audit_course, fix, and export_course. Limited to 10 imports per person (or API key) per 24 hours.

Input

{
  "type": "object",
  "properties": {
    "url": {
      "type": "string",
      "description": "Public https URL of the package. Give this or storagePath, not both."
    },
    "storagePath": {
      "type": "string",
      "description": "Path of a package already staged in course-imports storage (<org id>/<file>). Give this or url."
    },
    "filename": {
      "type": "string",
      "description": "File name with extension, e.g. \"safety.zip\" — needed when the URL or path does not end in one; the extension picks the parser"
    }
  },
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "import_course",
    "arguments": {
      "url": "https://example.com/legacy/forklift-safety-scorm12.zip"
    }
  }
}

Example result (structuredContent)

{
  "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "title": "Forklift Safety Refresher",
  "sourceFormat": "scorm12",
  "lessonCount": 6,
  "assetCount": 14,
  "report": {
    "notRecreated": [
      {
        "feature": "Flash interactions",
        "count": 2,
        "detail": "Rebuild them with native interactive blocks."
      }
    ],
    "notes": []
  },
  "warnings": [],
  "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
  "next": "Run audit_course on courseId, fix what it and report.notRecreated describe, then export_course."
}

Side effects

Creates a draft course from the package, uploads its media to the organization's media library, and deletes a staged storagePath file once read. Uses one of the 10 imports per 24 hours.

Errors

• Pass exactly one of url or storagePath.
• Cannot tell the package type from "<name>". Pass filename with one of these extensions: <extensions>.
• storagePath must be a file in your organization's course-imports folder (<org id>/<file>).
• No staged file at <storagePath>.
• Could not use that URL: <reason> It must be a public https URL.
• File too large (max 100MB)
• Importing a course needs the member role or higher in this organization.
• Import limit reached (<n> per 24 hours). Try again tomorrow.
• Failed to parse file: <parser message>
• Nothing in "<title>" could be converted into editable content, so no course was created. <reasons>

Notes

• The same pipeline as Courses → Import in the app, with the same limit: an OAuth connection shares the person's 10 imports per 24 hours with their imports in the app; an API key has its own budget. The limit refuses imports when it cannot be checked.
• report.notRecreated lists what the importer could not rebuild (with counts and advice); report.notes lists what it changed. Both are always arrays. Plan fixes from them, then run audit_course.
• The URL is fetched with the same guard as upload_asset: public https only, private and internal addresses refused at every redirect, at most 3 redirects, capped at 100MB.
• Storyline, Captivate and Lectora imports become slide-based (Scenes) courses: slideCount and objectCount are returned, and their content is edited in the app rather than with the block tools.
• AI modernization, offered by the app's importer, is not available over MCP.

scope write:courseswrites

Mint a shareable review link for a course. Anyone with the link sees the content read-only — block courses as lessons, scenes courses in the real player — and leaves structured feedback (typed comments, suggested replacement wording, per-block or per-slide anchors) without signing in. Hand the URL to the human to send to their reviewer; links expire after 30 days and can be revoked in the editor's Share dialog.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "title": {
      "type": "string",
      "description": "Optional label, e.g. \"Legal review — April\""
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "create_review_link",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "SME review — compliance pass"
    }
  }
}

Example result (structuredContent)

{
  "reviewUrl": "https://scormbusters.com/review/5a7b9c0d-1e2f-4a3b-8c4d-6e7f8091a2b3",
  "sessionId": "4b6c8d9e-0f1a-4b2c-9d3e-5f60718293a4",
  "title": "SME review — compliance pass",
  "expiresAt": "2026-11-01T09:00:00.000Z",
  "next": "Give reviewUrl to the human to send to their reviewer; later, read what came back with get_review_feedback."
}

Side effects

Creates a review link anyone can open without signing in (read-only content, comment-only). Expires after 30 days; revocable from the editor's Share dialog.

Errors

• Course <courseId> not found in your organization.
• Could not create the review link.

Notes

• The reviewer page shows block courses lesson by lesson and scenes courses in the real player, and collects structured feedback: a comment kind, an optional suggested replacement wording, and a per-block or per-slide anchor with a snapshot of what the reviewer was looking at.

scope read:coursesread-only

Everything reviewers have said about a course, across all of its review links: comments sorted most-urgent first with their kind (factual error, missing content, wording, …), the reviewer's suggested replacement text where given, whether the content has changed or been deleted since (context_state), reply threads, resolution state, and per-link verdicts (approved / changes requested). Start here before editing in response to a review; a comment whose context_state is "changed" may no longer apply.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    }
  },
  "required": [
    "courseId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_review_feedback",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11"
    }
  }
}

Example result (structuredContent)

{
  "sessions": [
    {
      "id": "4b6c8d9e-0f1a-4b2c-9d3e-5f60718293a4",
      "title": "SME review — compliance pass",
      "status": "changes_requested",
      "expires_at": "2026-11-01T09:00:00.000Z",
      "last_viewed_at": "2026-10-02T08:12:00.000Z",
      "created_at": "2026-10-01T09:00:00.000Z",
      "comment_count": 2
    }
  ],
  "comments": [
    {
      "id": "6c8d0e1f-2a3b-4c4d-ae5f-708192a3b4c5",
      "kind": "factual",
      "text": "The retention period is six years now, not five.",
      "suggested_text": "Keep every record for six years from the date of issue.",
      "author_name": "Dana (Legal)",
      "block_id": "f7410216-805d-4e9f-b142-cd3e6f708166",
      "lesson_id": "c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
      "lesson_title": "Introduction",
      "slide_id": null,
      "slide_title": null,
      "context_state": "current",
      "context_excerpt": "Keep every record for five years from the date of issue.",
      "resolved": false,
      "replies": [],
      "created_at": "2026-10-02T08:10:00.000Z"
    }
  ],
  "summary": {
    "total": 2,
    "actionable": 1,
    "approvals": 1,
    "questions": 0,
    "unclassified": 0,
    "byKind": {}
  },
  "unresolved_summary": {
    "total": 1,
    "actionable": 1,
    "approvals": 0,
    "questions": 0,
    "unclassified": 0,
    "byKind": {}
  }
}

Side effects

None.

Errors

• Course <courseId> not found in your organization.

Notes

• comments are sorted most-urgent first (factual errors before wording), threads nested under their first comment, and each carries context_state: current (safe to act on), changed (the content moved on — reread before acting), deleted, or unknown.
• suggested_text is the reviewer's replacement wording; when context_state is current, apply_review_suggestion lands it in one call.
• A session's status is the reviewer's overall verdict: open, changes_requested, or approved.

scope write:courseswritesdestructive

Accept a reviewer's suggested wording as written: replaces the text of the block the comment points at and marks the comment resolved. Refuses (409) when the block has been edited since the suggestion was written, when it was deleted, or when a concurrent edit lands first — never applies a suggestion onto words it was not about. On a refusal, read the comment and edit by hand with update_block instead.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "commentId": {
      "type": "string",
      "description": "The review comment id, from get_review_feedback"
    }
  },
  "required": [
    "courseId",
    "commentId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "apply_review_suggestion",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "commentId": "6c8d0e1f-2a3b-4c4d-ae5f-708192a3b4c5"
    }
  }
}

Example result (structuredContent)

{
  "blockId": "f7410216-805d-4e9f-b142-cd3e6f708166",
  "revision": 7,
  "resolved": true
}

Side effects

Replaces the anchored block's text with the reviewer's suggested wording (the block's revision advances, so a concurrent editor gets a conflict rather than losing work) and marks the comment resolved.

Errors

• Course <courseId> not found in your organization.
• Comment not found
• This comment has no suggested wording
• The block this suggestion was written for no longer exists
• The block has been edited since this suggestion was written — review it by hand instead
• This block has no text field the suggestion could replace
• The block changed while applying — try again

Notes

• Only applies onto the exact words the reviewer was reading (a stored content hash is checked server-side). On any refusal, read the comment and make the edit yourself with update_block, then resolve_review_comment.

scope write:courseswritesdestructive

Mark a reviewer's comment resolved (or reopen it). Resolve after the change it asked for has actually been made — the reviewer sees the resolution on their page, so a false resolve burns trust.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "commentId": {
      "type": "string",
      "description": "The review comment id, from get_review_feedback"
    },
    "resolved": {
      "type": "boolean",
      "description": "true to resolve, false to reopen"
    }
  },
  "required": [
    "courseId",
    "commentId",
    "resolved"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "resolve_review_comment",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "commentId": "6c8d0e1f-2a3b-4c4d-ae5f-708192a3b4c5",
      "resolved": true
    }
  }
}

Example result (structuredContent)

{
  "commentId": "6c8d0e1f-2a3b-4c4d-ae5f-708192a3b4c5",
  "resolved": true
}

Side effects

Sets the comment's resolved flag; the reviewer sees the resolution on their page.

Errors

• Course <courseId> not found in your organization.
• Comment not found
• Could not update the comment.

Notes

• Resolve only after the change it asked for was actually made.

scope write:courseswrites

Post a reply in a review comment's thread. The reviewer sees it on their review page, labeled as coming from the author's side — use it to ask what a comment means or to say what was changed, instead of letting the question leak into email. Keep it short and concrete.

Input

{
  "type": "object",
  "properties": {
    "courseId": {
      "type": "string",
      "description": "Course id (uuid)"
    },
    "commentId": {
      "type": "string",
      "description": "The thread's first comment id, from get_review_feedback"
    },
    "text": {
      "type": "string",
      "description": "The reply, plain text, max 5000 characters"
    }
  },
  "required": [
    "courseId",
    "commentId",
    "text"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "reply_review_comment",
    "arguments": {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "commentId": "6c8d0e1f-2a3b-4c4d-ae5f-708192a3b4c5",
      "text": "Updated to six years — thanks for catching it."
    }
  }
}

Example result (structuredContent)

{
  "replyId": "7d9e1f20-3b4c-4d5e-bf60-8192a3b4c5d6",
  "createdAt": "2026-10-02T09:30:00.000Z"
}

Side effects

Posts a reply into the comment's thread, visible on the reviewer's page and labeled as coming from the author's side.

Errors

• Course <courseId> not found in your organization.
• Comment not found
• Reply text is required.
• Reply must be 5000 characters or fewer.
• Reply to the thread's first comment, not to a reply.
• Could not post the reply.

Notes

• Over an OAuth connection the reply carries the person's name; over an API key it is posted as "The author's assistant".

scope read:coursesread-only

Enterprise. Find text across every course in the organization: block content (paragraphs, questions, options, feedback, captions, alt text…), Scenes slides (text objects, states, alt text, slide and layer names, notes) and course, module and lesson titles. Returns one row per occurrence with context, whether it can be replaced automatically, and an editorUrl that opens the editor at it. Use it to answer "where do we still say X?" before deciding on a replace_in_library.

Input

{
  "type": "object",
  "properties": {
    "find": {
      "type": "string",
      "description": "Text to find (1–200 characters), or a pattern with regex: true"
    },
    "matchCase": {
      "type": "boolean",
      "description": "Match upper and lower case exactly (default false)"
    },
    "wholeWord": {
      "type": "boolean",
      "description": "Only whole words (default false)"
    },
    "regex": {
      "type": "boolean",
      "description": "Treat find as a JavaScript regular expression (default false). Nested repetition like (a+)+ and backreferences are refused"
    },
    "courseIds": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Only these courses (uuids)"
    },
    "statuses": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "draft",
          "published",
          "archived"
        ]
      },
      "description": "Only courses with these statuses"
    },
    "sources": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Only courses imported from these tools (e.g. \"storyline\", \"rise\"); \"none\" for courses built in Scormbusters"
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 500,
      "description": "Most occurrences to list (default 100); totals always count every match"
    }
  },
  "required": [
    "find"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "search_library",
    "arguments": {
      "find": "five years",
      "wholeWord": true,
      "statuses": [
        "published"
      ]
    }
  }
}

Example result (structuredContent)

{
  "totals": {
    "courses": 1,
    "occurrences": 2,
    "replaceable": 1,
    "needsManualEdit": 1,
    "coursesSearched": 12,
    "coursesInScope": 12,
    "truncated": false,
    "partial": false,
    "skippedTooLong": 0
  },
  "courses": [
    {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Forklift Safety Refresher",
      "status": "published",
      "source": null,
      "occurrences": 2
    }
  ],
  "matches": [
    {
      "occurrenceId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11|block|f7410216-805d-4e9f-b142-cd3e6f708166|content:text|0|8e3f12a4",
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "courseTitle": "Forklift Safety Refresher",
      "where": "The daily inspection › Heading › Text",
      "before": "Keep every record for ",
      "match": "five years",
      "after": " from the date of issue.",
      "replaceable": true,
      "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11?lesson=c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33&block=f7410216-805d-4e9f-b142-cd3e6f708166"
    },
    {
      "occurrenceId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11|block|19632438-a27f-4ab1-9364-ef508192a388|content:question|0|5b1d07e2",
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "courseTitle": "Forklift Safety Refresher",
      "where": "Check your knowledge › Multiple Choice › Question",
      "before": "How long do we keep records? (",
      "match": "five years",
      "after": ")",
      "replaceable": false,
      "reason": "spans-formatting",
      "editorUrl": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11?lesson=d52f8094-6e3b-4c7d-9f20-ab1c4d5e6f44&block=19632438-a27f-4ab1-9364-ef508192a388"
    }
  ]
}

Side effects

None.

Errors

• Enter something to find.
• Search text must be <n> characters or fewer.
• That pattern is not a valid regular expression: <reason>
• That pattern repeats a group that itself repeats or alternates (like (a+)+ or (a|b)*), which can take minutes to run. Simplify it.
• Backreferences are not supported in library search.
• <field> must be a list of valid values.
• Library search and replace need the member role or higher in this organization.
• Library search and bulk replace — finding and changing text across every course at once — is an Enterprise feature. Find and edit within a course on every plan.

Notes

• Searched: every learner-visible field of every block type (the same fields translation covers — never answer keys, URLs or ids), Scenes text objects and their state overrides, question text, options, feedback, alt text, accessible names, captions, slide names and notes, layer and object names, and course, module and lesson titles and descriptions.
• Rich text is matched as the learner reads it: tags skipped, entities decoded, bold or linked words joined. A match that crosses formatting has replaceable: false and reason spans-formatting — edit it by hand at its editorUrl.
• totals count every match; matches lists at most limit of them. partial: true means the search stopped at its time budget before every course was read — narrow it with courseIds, statuses or sources.

scope write:courseswritesdestructive

Enterprise. Replace text across many courses at once, in two calls. First call with dry_run: true — nothing is written; you get the before/after preview of every occurrence and a confirm_token. Show the human the preview. Only if they approve, call again with the same arguments, dry_run: false and the confirm_token. Rich text is only changed inside one run of text: a match that crosses formatting (part bold) is reported as needing a manual edit, never guessed at. Every changed lesson or Scenes course gets a version-history snapshot first, each course is audited, and the whole edit can be reversed with undo_library_replace.

Input

{
  "type": "object",
  "properties": {
    "find": {
      "type": "string",
      "description": "Text to find (1–200 characters), or a pattern with regex: true"
    },
    "matchCase": {
      "type": "boolean",
      "description": "Match upper and lower case exactly (default false)"
    },
    "wholeWord": {
      "type": "boolean",
      "description": "Only whole words (default false)"
    },
    "regex": {
      "type": "boolean",
      "description": "Treat find as a JavaScript regular expression (default false). Nested repetition like (a+)+ and backreferences are refused"
    },
    "courseIds": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Only these courses (uuids)"
    },
    "statuses": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "draft",
          "published",
          "archived"
        ]
      },
      "description": "Only courses with these statuses"
    },
    "sources": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Only courses imported from these tools (e.g. \"storyline\", \"rise\"); \"none\" for courses built in Scormbusters"
    },
    "replace": {
      "type": "string",
      "description": "Replacement text (may be empty to delete the matches). With regex: true, $1…$9 and $<name> insert groups"
    },
    "dry_run": {
      "type": "boolean",
      "description": "true to preview without writing (required first); false to apply"
    },
    "confirm_token": {
      "type": "string",
      "description": "The confirm_token from the dry run; required with dry_run: false"
    },
    "occurrenceIds": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Optional: replace only these occurrences (ids from the preview). Omit to replace every replaceable occurrence"
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 200,
      "description": "Most preview rows to list (default 50)"
    }
  },
  "required": [
    "find",
    "replace",
    "dry_run"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "replace_in_library",
    "arguments": {
      "find": "five years",
      "replace": "six years",
      "wholeWord": true,
      "statuses": [
        "published"
      ],
      "dry_run": false,
      "confirm_token": "9c1e4f0a7b2d4c6e8f10a2b3c4d5e6f7"
    }
  }
}

Example result (structuredContent)

{
  "dry_run": false,
  "editId": "2a4c6e80-9b1d-4f3a-8c5e-7d9f0b1a2c3d",
  "totals": {
    "courses": 1,
    "replaced": 1,
    "skipped": 0
  },
  "courses": [
    {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Forklift Safety Refresher",
      "replaced": 1,
      "skipped": 0
    }
  ],
  "next": "Every changed lesson and Scenes course has a version-history snapshot labeled with this replace. To reverse the whole edit, call undo_library_replace with editId."
}

Side effects

dry_run: true writes nothing. dry_run: false first saves a version-history snapshot of every lesson and Scenes course it will change (labeled "Bulk replace: '<find>' → '<replace>'"), then rewrites the matching text, writes one audit-log entry per changed course, records the edit for undo_library_replace, and updates any editor that has the course open.

Errors

• dry_run is required: call with dry_run: true first, check the preview with the human, then repeat with dry_run: false and the confirm_token.
• confirm_token is required with dry_run: false. Run a dry run first and pass the confirm_token it returns.
• The library changed since the dry run, or the arguments differ from it. Run replace_in_library with dry_run: true again and confirm the new preview.
• replace must be a string (it may be empty, to delete the matches).
• replace must be <n> characters or fewer.
• Bulk edits are limited to <n> an hour. Try again later.
• Library search and replace need the member role or higher in this organization.
• Enter something to find.

Notes

• The confirm_token is a digest of the arguments and of exactly which occurrences the dry run previewed. If the library changes in between (someone edits a matching field) or any argument differs, the apply is refused and nothing is written: dry-run again.
• A block is only written at the revision the replace read; an author's save that lands in between wins and the occurrence is reported as edited-concurrently. A course too large to snapshot is left unchanged, with an error naming why.
• Inserted text is escaped in rich text, so a replacement can never add markup. A fill-in-the-blank sentence whose [blank] count would change is skipped (blank-mismatch).
• No confirm_token is issued when the search was cut short (more than 5000 matches, or it ran out of time): narrow it with courseIds, statuses or sources.

scope write:courseswritesdestructive

Enterprise. Reverse one replace_in_library: every field it changed goes back to what it was, wherever the text is still what the replace wrote. Fields someone edited since are left alone and listed as conflicts. Only the person who made the edit, or an admin, can undo it.

Input

{
  "type": "object",
  "properties": {
    "editId": {
      "type": "string",
      "description": "The editId replace_in_library returned"
    }
  },
  "required": [
    "editId"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "undo_library_replace",
    "arguments": {
      "editId": "2a4c6e80-9b1d-4f3a-8c5e-7d9f0b1a2c3d"
    }
  }
}

Example result (structuredContent)

{
  "editId": "2a4c6e80-9b1d-4f3a-8c5e-7d9f0b1a2c3d",
  "restored": 1,
  "courses": [
    {
      "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Forklift Safety Refresher",
      "restored": 1,
      "conflicts": 0
    }
  ],
  "conflicts": [],
  "next": "The bulk edit is fully undone."
}

Side effects

Saves a "Before undo" version-history snapshot of each affected lesson and Scenes course, puts every changed field back to its value before the replace where it still holds what the replace wrote, audits each course, and marks the edit undone.

Errors

• editId must be the id replace_in_library returned.
• Bulk edit not found in your organization.
• This bulk edit has already been undone.
• Only the person who made this bulk edit, or an admin, can undo it.
• Bulk edits are limited to <n> an hour. Try again later.

Notes

• Fields edited after the replace are not touched and come back in conflicts (reason edited-since, or deleted); restore them from the course's version history if needed.

scope read:coursesread-only

Search the words in the organization's courses — titles, descriptions, lesson text, questions and Scenes slide text — and return up to 20 matching course outlines, lessons and slides, best first: {results: [{id, title, url}]}. Read one with fetch(id).

Input

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "description": "Words to look for, e.g. \"fire extinguisher types\""
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "search",
    "arguments": {
      "query": "daily forklift inspection"
    }
  }
}

Example result (structuredContent)

{
  "results": [
    {
      "id": "lesson:3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11:c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
      "title": "Forklift Safety Refresher › The daily inspection",
      "url": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11?lesson=c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33"
    },
    {
      "id": "course:3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
      "title": "Forklift Safety Refresher",
      "url": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11"
    }
  ]
}

Side effects

None.

Errors

• query is required.

Notes

• The shape ChatGPT connectors and deep research expect; any MCP client can use it. Results are course outlines (course:<id>), lessons (lesson:<courseId>:<lessonId>) and Scenes slides (slide:<courseId>:<slideId>), ranked by how many of the query's words they contain.
• Searches every course in the organization, reading the courses whose titles match first; if a very large library runs past the time budget, the rest are matched on title and description only.
• For exact phrases, regular expressions and every occurrence with context, use search_library (Enterprise).

scope read:coursesread-only

Read the text of one search result — a whole course, a lesson or a slide — as {id, title, text, url, metadata}. A course is capped at 50,000 characters; fetch its lessons for the rest.

Input

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "An id from search, e.g. \"course:<uuid>\" or \"lesson:<courseId>:<lessonId>\""
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "fetch",
    "arguments": {
      "id": "lesson:3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11:c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33"
    }
  }
}

Example result (structuredContent)

{
  "id": "lesson:3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11:c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
  "title": "Forklift Safety Refresher › The daily inspection",
  "text": "# Forklift Safety Refresher › The daily inspection\nLesson · Lesson title: The daily inspection\nHeading · Text: Before you start",
  "url": "https://scormbusters.com/courses/3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11?lesson=c41e7f83-5d2a-4b6c-8e1f-9a0b3c4d5e33",
  "metadata": {
    "kind": "lesson",
    "courseId": "3f6c1a52-7a0e-4f4b-9d2c-5b8e1f0a9c11",
    "courseTitle": "Forklift Safety Refresher",
    "status": "draft",
    "editorMode": "classic",
    "updatedAt": "2026-09-25T10:00:00.000Z",
    "truncated": false
  }
}

Side effects

None.

Errors

• Unknown id "<id>". Pass an id exactly as search returned it.
• Nothing with id <id> in your organization.

Notes

• text is plain text, one line per field (markup removed), labeled with the block or object it came from.