Skip to main content
Glama

Canva

by com.canva.mcp
OverviewDiscussionsAdmin

Create Video

create-video

Produces a Canva video from a complete production specification.

Use when a production guide has been selected and its required inputs and assets are available.

Only call after the user has approved this video generation. Before the first generation call in a conversation, tell the user: "Just a quick heads up - creating a video may use a significant portion or all of your monthly AI limit. Would you like to continue?" Wait for their explicit agreement. For later calls, do not repeat the heads-up, but require the user to request or approve each new video generation.

Refer to it as their "AI limit". Do not provide their remaining AI limit, estimate how much a video will use, or discuss prices, plan names, or upgrades.

Returns a job_id and polling_policy for tracking generation to completion.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
user_intentNoMandatory description of what the user is trying to accomplish with this tool call. This should always be provided by LLM clients. Please keep it concise (255 characters or less recommended).
video_titleYesInfer and provide a concise, descriptive video name based on the user's request, for example "Summer sale teaser". Do not require the user to supply an exact title. Plain text: letters, numbers, spaces and ,.'!?&:()-_/#%+@$*;=|~[]’‘“”–—" only. No emoji, rare symbols, backslashes, angle brackets, line breaks or file extension. Max 250 characters.
production_specYesComplete production plan following the selected production guide, including the brief, output settings, visual direction, storyboard, required assets, audio, branding, constraints, and disclosed assumptions. Total serialized UTF-8 size <= 64 KiB.
video_generation_confirmedYesSet to true only when the user has approved this video generation. Before the first generation call in a conversation, approval must follow this heads-up: "Just a quick heads up - creating a video may use a significant portion or all of your monthly AI limit. Would you like to continue?" For later calls, do not repeat the heads-up, but require the user to request or approve each new video generation. For retries after a failed job, do not repeat the AI-limit notice. The user's request to retry counts as approval; never retry automatically.

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
job_idYesID of the video generation job. Pass it to get-video-generation-status to poll for the result. This value is stable and never changes.
statusYesStatus of the job. It has only just been started, so poll get-video-generation-status with the job_id for progress and the result.
polling_policyYesHow and when to poll next.
continuation_tokenYesOpaque token to pass on the first poll to get-video-generation-status. Always use the most recent token from the last response.
expected_duration_secondsNoEstimated total generation time in seconds, when available.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changed
    • changedOutput schema / properties / polling_policy / properties / instructions / description
      Previous value: -"Instructions for polling this job and communicating what happens next. They MUST be followed. Relay the initial generation-time expectation and Canva Projects link to the user, then continue polling as instructed while polling_policy is present."New value: +"Instructions for polling this job and communicating what happens next. They MUST be followed. Relay exactly once that video generation can take up to 20 minutes and will appear at the provided Canva Projects link only after generation finishes. Continue polling as instructed while polling_policy is present; do not treat the Projects link as evidence that the video is ready."
  2. Changed1 schema field changed
    • changedInput schema / properties / video_title / description
      Previous value: -"Always provide a video title from the user's request, e.g. \"Summer sale teaser\". Plain text: letters, numbers, spaces and ,.'!?&:()-_/#%+@$*;=|~[]’‘“”–—\" only. No emoji, rare symbols, backslashes, angle brackets, line breaks or file extension. Max 250 characters. An invalid title is ignored."New value: +"Infer and provide a concise, descriptive video name based on the user's request, for example \"Summer sale teaser\". Do not require the user to supply an exact title. Plain text: letters, numbers, spaces and ,.'!?&:()-_/#%+@$*;=|~[]’‘“”–—\" only. No emoji, rare symbols, backslashes, angle brackets, line breaks or file extension. Max 250 characters."
  3. Changed2 schema fields changed
    • addedInput schema / properties / video_title
      Added value: +{
+  "description": "Always provide a video title from the user's request, e.g. \"Summer sale teaser\". Plain text: letters, numbers, spaces and ,.'!?&:()-_/#%+@$*;=|~[]’‘“”–—\" only. No emoji, rare symbols, backslashes, angle brackets, line breaks or file extension. Max 250 characters. An invalid title is ignored.",
+  "maxLength": 250,
+  "type": "string"
+}
    • changedInput schema / required
      Previous value: -[
-  "video_generation_confirmed",
-  "production_spec"
-]New value: +[
+  "video_generation_confirmed",
+  "production_spec",
+  "video_title"
+]
  4. Changed3 schema fields changed
    • addedOutput schema / properties / continuation_token
      Added value: +{
+  "description": "Opaque token to pass on the first poll to get-video-generation-status. Always use the most recent token from the last response.",
+  "type": "string"
+}
    • changedOutput schema / properties / polling_policy / properties / instructions / description
      Previous value: -"Instructions for polling this job and communicating what happens next. They MUST be followed. Relay the initial generation-time expectation and Canva Uploads link to the user, then continue polling as instructed while polling_policy is present."New value: +"Instructions for polling this job and communicating what happens next. They MUST be followed. Relay the initial generation-time expectation and Canva Projects link to the user, then continue polling as instructed while polling_policy is present."
    • changedOutput schema / required
      Previous value: -[
-  "job_id",
-  "status",
-  "polling_policy"
-]New value: +[
+  "job_id",
+  "status",
+  "polling_policy",
+  "continuation_token"
+]
  5. Changed6 schema fields changed
    • changedInput schema / properties / production_spec / description
      Previous value: -"A complete consumer-authored production plan. The consuming assistant must do the directing work, not delegate a vague prompt to another writer. Total serialized UTF-8 size <= 64 KiB."New value: +"Complete production plan following the selected production guide, including the brief, output settings, visual direction, storyboard, required assets, audio, branding, constraints, and disclosed assumptions. Total serialized UTF-8 size <= 64 KiB."
    • changedInput schema / properties / production_spec / properties / audio / anyOf
      Previous value: -[
-  {
-    "additionalProperties": false,
-    "description": "No spoken dialogue.",
-    "properties": {
-      "ambience": {
-        "description": "Diegetic sound and specific action-synced effects; use none for silence.",
-        "maxLength": 350,
-        "minLength": 1,
-        "type": "string"
-      },
-      "mode": {
-        "const": "none",
-        "type": "string"
-      },
-      "music": {
-        "description": "Music mood, instrumentation, tempo and relationship to cuts; use none for no music.",
-        "maxLength": 350,
-        "minLength": 1,
-        "type": "string"
-      }
-    },
-    "required": [
-      "mode",
-      "music",
-      "ambience"
-    ],
-    "type": "object"
-  },
-  {
-    "additionalProperties": false,
-    "description": "Generated speech with an explicitly directed voice.",
-    "properties": {
-      "accent": {
-        "description": "Requested regional accent or neutral; do not imitate a named real person.",
-        "maxLength": 100,
-        "minLength": 1,
-        "type": "string"
-      },
-      "ambience": {
-        "description": "Soundscape and action-synced effects; use none for silence.",
-        "maxLength": 350,
-        "minLength": 1,
-        "type": "string"
-      },
-      "language": {
-        "description": "Language of the exact supplied script; never silently translate it.",
-        "maxLength": 60,
-        "minLength": 1,
-        "type": "string"
-      },
-      "mode": {
-        "description": "On-camera requires a presenter reference; voice-over is not visible lip-sync.",
-        "enum": [
-          "voice-over",
-          "on-camera"
-        ],
-        "type": "string"
-      },
-      "music": {
-        "description": "Music mood, instrumentation, tempo; keep below speech or use none.",
-        "maxLength": 350,
-        "minLength": 1,
-        "type": "string"
-      },
-      "script": {
-        "description": "Exact approved words, equal to nonempty storyboard dialogue joined with newline. Fit comfortably in the duration; no invented factual claims.",
-        "maxLength": 600,
-        "minLength": 1,
-        "type": "string"
-      },
-      "voice_direction": {
-        "description": "Vocal age range, texture, pace, energy and pauses, without naming a real person's voice.",
-        "maxLength": 300,
-        "minLength": 1,
-        "type": "string"
-      }
-    },
-    "required": [
-      "mode",
-      "script",
-      "language",
-      "accent",
-      "voice_direction",
-      "music",
-      "ambience"
-    ],
-    "type": "object"
-  }
-]New value: +[
+  {
+    "additionalProperties": false,
+    "description": "No spoken dialogue.",
+    "properties": {
+      "ambience": {
+        "description": "Diegetic sound and specific action-synced effects; use none for silence.",
+        "maxLength": 350,
+        "minLength": 1,
+        "type": "string"
+      },
+      "mode": {
+        "const": "none",
+        "type": "string"
+      },
+      "music": {
+        "description": "Music mood, instrumentation, tempo and relationship to cuts; use none for no music.",
+        "maxLength": 350,
+        "minLength": 1,
+        "type": "string"
+      }
+    },
+    "required": [
+      "mode",
+      "music",
+      "ambience"
+    ],
+    "type": "object"
+  },
+  {
+    "additionalProperties": false,
+    "description": "Generated speech with an explicitly directed voice.",
+    "properties": {
+      "accent": {
+        "description": "Requested regional accent or neutral; do not imitate a named real person.",
+        "maxLength": 100,
+        "minLength": 1,
+        "type": "string"
+      },
+      "ambience": {
+        "description": "Soundscape and action-synced effects; use none for silence.",
+        "maxLength": 350,
+        "minLength": 1,
+        "type": "string"
+      },
+      "language": {
+        "description": "Language of the exact supplied script; never silently translate it.",
+        "maxLength": 60,
+        "minLength": 1,
+        "type": "string"
+      },
+      "mode": {
+        "description": "On-camera requires a presenter reference; voice-over is not visible lip-sync.",
+        "enum": [
+          "voice-over",
+          "on-camera"
+        ],
+        "type": "string"
+      },
+      "music": {
+        "description": "Music mood, instrumentation, tempo; keep below speech or use none.",
+        "maxLength": 350,
+        "minLength": 1,
+        "type": "string"
+      },
+      "script": {
+        "description": "Exact approved words, equal to nonempty storyboard dialogue joined with newlines. Allow enough time for natural delivery, using no more than three words per second. Do not invent factual claims.",
+        "maxLength": 600,
+        "minLength": 1,
+        "type": "string"
+      },
+      "voice_direction": {
+        "description": "Vocal age range, texture, pace, energy and pauses, without naming a real person's voice.",
+        "maxLength": 300,
+        "minLength": 1,
+        "type": "string"
+      }
+    },
+    "required": [
+      "mode",
+      "script",
+      "language",
+      "accent",
+      "voice_direction",
+      "music",
+      "ambience"
+    ],
+    "type": "object"
+  }
+]
    • changedInput schema / properties / production_spec / properties / reference_images / items / properties / source / anyOf
      Previous value: -[
-  {
-    "additionalProperties": false,
-    "description": "A genuine authorized Canva image.",
-    "properties": {
-      "asset_id": {
-        "description": "Real ID returned by get-assets or an upload tool. No placeholders, URLs or paths.",
-        "maxLength": 50,
-        "minLength": 1,
-        "pattern": "^[A-Za-z0-9_-]+$",
-        "type": "string"
-      },
-      "kind": {
-        "const": "canva_asset",
-        "type": "string"
-      }
-    },
-    "required": [
-      "kind",
-      "asset_id"
-    ],
-    "type": "object"
-  },
-  {
-    "additionalProperties": false,
-    "description": "An intentionally fictional reference to generate before rendering.",
-    "properties": {
-      "image_prompt": {
-        "description": "Production-ready image direction: subject identity, full silhouette, neutral background where appropriate, light, angle, materials and no unintended lettering.",
-        "maxLength": 1200,
-        "minLength": 1,
-        "type": "string"
-      },
-      "kind": {
-        "const": "generate",
-        "type": "string"
-      }
-    },
-    "required": [
-      "kind",
-      "image_prompt"
-    ],
-    "type": "object"
-  }
-]New value: +[
+  {
+    "additionalProperties": false,
+    "description": "A genuine authorized Canva image.",
+    "properties": {
+      "asset_id": {
+        "description": "Real Canva asset ID for the user-provided image. If the image was attached in chat, resolve or upload the attachment first and use the returned asset ID. Do not ask the user to provide the same image again, and never invent an asset ID.",
+        "maxLength": 50,
+        "minLength": 1,
+        "pattern": "^[A-Za-z0-9_-]+$",
+        "type": "string"
+      },
+      "kind": {
+        "const": "canva_asset",
+        "type": "string"
+      }
+    },
+    "required": [
+      "kind",
+      "asset_id"
+    ],
+    "type": "object"
+  },
+  {
+    "additionalProperties": false,
+    "description": "An intentionally fictional reference to generate before rendering.",
+    "properties": {
+      "image_prompt": {
+        "description": "Production-ready image direction: subject identity, full silhouette, neutral background where appropriate, light, angle, materials and no unintended lettering.",
+        "maxLength": 1200,
+        "minLength": 1,
+        "type": "string"
+      },
+      "kind": {
+        "const": "generate",
+        "type": "string"
+      }
+    },
+    "required": [
+      "kind",
+      "image_prompt"
+    ],
+    "type": "object"
+  }
+]
    • addedInput schema / properties / video_generation_confirmed
      Added value: +{
+  "description": "Set to true only when the user has approved this video generation. Before the first generation call in a conversation, approval must follow this heads-up: \"Just a quick heads up - creating a video may use a significant portion or all of your monthly AI limit. Would you like to continue?\" For later calls, do not repeat the heads-up, but require the user to request or approve each new video generation. For retries after a failed job, do not repeat the AI-limit notice. The user's request to retry counts as approval; never retry automatically.",
+  "type": "boolean"
+}
    • changedInput schema / required
      Previous value: -[
-  "production_spec"
-]New value: +[
+  "video_generation_confirmed",
+  "production_spec"
+]
    • changedOutput schema / properties / polling_policy / properties / instructions / description
      Previous value: -"Instructions on when to poll this job next, in words. They MUST be followed. Relay them to the user while they wait."New value: +"Instructions for polling this job and communicating what happens next. They MUST be followed. Relay the initial generation-time expectation and Canva Uploads link to the user, then continue polling as instructed while polling_policy is present."
  6. Changed3 schema fields changed
    • changedInput schema / properties / production_spec / properties / output / properties / engine / enum
      Previous value: -[
-  "seedance",
-  "h3"
-]New value: +[
+  "seedance"
+]
    • changedInput schema / properties / production_spec / properties / output / properties / resolution / description
      Previous value: -"Exact engine-specific value: seedance=720p; h3=768P."New value: +"Seedance production resolution: 720p."
    • changedInput schema / properties / production_spec / properties / output / properties / resolution / enum
      Previous value: -[
-  "720p",
-  "768P"
-]New value: +[
+  "720p"
+]
  7. Added

TDQS

A4.4/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare the safety profile (not read-only, openWorld, non-destructive), and the description adds substantial behavior beyond that: the AI-limit consumption heads-up, the per-call approval requirement, forbidden topics (remaining limit, estimates, prices, upgrades), and the returned job_id/polling_policy. This is unusually rich behavioral disclosure.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with purpose, then usage, then user-facing policy, then return values. It is longer than typical but each block earns its place against the tool's complexity; some wording is duplicated in the confirmation parameter description, a minor redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a high-complexity creation tool with a deep nested schema, existing annotations, and an output schema, the description covers the preconditions (guide selected, assets ready), the approval gate, and the async follow-up, which is everything an agent needs to call it correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and the schema already documents every field exhaustively (production_spec, video_title, video_generation_confirmed). The description adds no syntax or semantic detail beyond the schema, so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ("Produces a Canva video") and names the input artifact ("a complete production specification"). The precondition "when a production guide has been selected" implicitly separates it from siblings like generate-design and generate-image, but no sibling is named directly, so differentiation is contextual rather than explicit.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives a concrete when-to-use condition ("a production guide has been selected and its required inputs and assets are available") and a hard when-not gate ("Only call after the user has approved"). It spells out the exact approval and disclosure flow, including the first-call vs later-call distinction, leaving nothing to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Canva toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources