Opus Motion

Make launch and promo videos in a brand from designed templates: search, draft, edit and preview.

使うべきか

品質と安全性

A
説明の品質
100%
スキーマの完全性
94%
命名の品質
93%
ポイズニングのリスク
80%
権限の一致
100%
プロトコルへの準拠
100%

検出事項(2)

  • HIGHTool poisoning patterns detected
  • INFOTool description contains placeholder or incomplete textjoin_studio_session 内

ツール定義とプロトコルへの準拠に関する自動分析に基づいています。

コンテキストコスト

~4,559トークン数(ツール定義)
~1.9 KB一般的なレスポンスサイズ
注意への影響は大きい(128k コンテキストの 3.56%)

これは、サーバーのツールがモデルのコンテキストに読み込まれるたびに消費されるおおよそのトークン数です。数が多いほど、ほかのタスクに使える注意が減ります。

インストール

ワンクリックインストール

これを `claude_desktop_config.json` ファイルに追加してください:

{
  "mcpServers": {
    "opus-motion": {
      "url": "https://opusmotion.com/api/mcp"
    }
  }
}

リモートエンドポイント

https://opusmotion.com/api/mcpstreamable-http

できること

ツール一覧

ツール(17)

🟢 読み取り専用🟡 書き込み🔴 削除⚪ 不明
🟢search_templates(query, category, tier, format)

Search the Opus Motion catalog of promo and launch video templates. Returns each template's id, title, description, tier and price, length, formats and example brands. Use it to find a template for someone's launch, product or brand video. Pro templates (every layer editable) are listed by list_pro_templates.

入力スキーマ

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "description": "One or two words: a use, a style or a brand (for example 'launch' or 'music app'). Leave it out to see every template."
    },
    "category": {
      "type": "string",
      "description": "A category id, for example 'launches'."
    },
    "tier": {
      "type": "string",
      "enum": [
        "free",
        "premium"
      ],
      "description": "Only free or only premium templates."
    },
    "format": {
      "type": "string",
      "description": "Only templates offered in this format."
    }
  },
  "additionalProperties": false
}
🟢get_template(id, images)

Everything about one catalog template: description, tier and price, length, formats, its scenes (what each shows, when), every field it takes (texts with their limits, images with their shape, colors, choices, clips) and, when it has them, the safe adjustments (tweaks). With images: true, its timelapse (the whole video in one picture) comes back inside the answer.

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The template's id (t07-discover-play) or slug (discover-play)."
    },
    "images": {
      "type": "boolean",
      "description": "Also return the template's timelapse image inside the answer."
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
🟢studio_link(template, site, format)

A link the person opens to see the template made with their brand: the studio reads their website (logo, colors, copy) and fills the template. Give it to the person; exporting needs their own account, and payment for premium templates.

入力スキーマ

{
  "type": "object",
  "properties": {
    "template": {
      "type": "string",
      "description": "The template's id or slug."
    },
    "site": {
      "type": "string",
      "description": "The person's website domain, for example example.com."
    },
    "format": {
      "type": "string",
      "description": "Open the studio in this format."
    }
  },
  "required": [
    "template"
  ],
  "additionalProperties": false
}
🟡join_studio_session(code)

The person has a template open in the Opus Motion studio and gave you its session code (OM-XXXX-XXXX). Join it first: you get the draft their studio shows, its editToken and the next steps. Then change it with update_draft: each change appears in their studio within seconds, and what they change there is in the draft when you read it. Never make a new draft for a session, and don't hand over a link: they have it open.

入力スキーマ

{
  "type": "object",
  "properties": {
    "code": {
      "type": "string",
      "description": "The session code from the person's prompt, for example OM-7KQ2-X4PD."
    }
  },
  "required": [
    "code"
  ],
  "additionalProperties": false
}
🟡create_draft(template, format, site, brand, values, ...)

Make a draft of a catalog template with the person's texts, images, colors and safe size adjustments. Returns a studioUrl the person opens to see the video playing live, an editToken to change it later (update_draft) and to look at it (look_at_draft), what is missing and the price to tell them. Nothing is rendered; the person exports it.

入力スキーマ

{
  "type": "object",
  "properties": {
    "template": {
      "type": "string",
      "description": "The template's id or slug (from search_templates)."
    },
    "format": {
      "type": "string",
      "description": "The format the draft opens in."
    },
    "site": {
      "type": "string",
      "description": "The person's website domain, for example example.com: texts you leave out are filled from it when it was read before."
    },
    "brand": {
      "type": "object",
      "description": "The brand: name, colors (1 to 4 hex) and logo or icon (base64 data URLs, 1 MB each)."
    },
    "values": {
      "type": "object",
      "description": "Field id → value: texts, choices, colors, and small images as base64 data URLs. get_template lists the fields. Image files and clips go with media_upload_command."
    },
    "tweaks": {
      "type": "object",
      "description": "Safe adjustments the template declares, for example { \"logoScale\": 1.3 }."
    }
  },
  "required": [
    "template"
  ],
  "additionalProperties": false
}
🟡update_draft(id, editToken, format, site, brand, ...)

Change a draft made with create_draft (the same link stays): send only what changes; null takes a field back to the template's.

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "editToken": {
      "type": "string",
      "description": "The editToken the draft came with."
    },
    "format": {
      "type": "string",
      "description": "A format id the template offers, for example 16:9, 9:16 or 1:1."
    },
    "site": {
      "type": "string"
    },
    "brand": {
      "type": "object"
    },
    "values": {
      "type": "object"
    },
    "tweaks": {
      "type": "object"
    }
  },
  "required": [
    "id",
    "editToken"
  ],
  "additionalProperties": false
}
🟢get_draft(id)

What a draft holds: its template, format, values, what is missing, its studio link and its price.

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The draft's id (draftId)."
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
🟢media_upload_command(id, field, file, editToken)

An image file (PNG, JPEG, WebP or SVG, up to 4 MB) or a clip (MP4 or MOV, up to 40 MB) is too big to send as a tool argument: this answers the exact HTTP request, and a curl command, that puts the file into one of the draft's image or clip fields, checked like any value. Run it from your shell with the file's path.

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "field": {
      "type": "string",
      "description": "An image or video field (get_template lists them)."
    },
    "file": {
      "type": "string",
      "description": "The file's path, to write it into the command (its extension says its type)."
    },
    "editToken": {
      "type": "string",
      "description": "The draft's editToken, to write it into the command (left out, the command has a placeholder)."
    }
  },
  "required": [
    "id",
    "field"
  ],
  "additionalProperties": false
}
🟢list_pro_templates

The Pro templates: designed videos in 16:9 and 9:16 whose every text, image and color an agent can change for a brand. Each lists its length and formats.

入力スキーマ

{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
🟢get_pro_template(id)

Everything to put a brand on a Pro template: its summary and acts, its palette by role (one recolor per role makes it the brand's), and every layer: what it is for and how long its text should be (note), when it shows, how it looks and moves, and its box and size in each format (read-only: the layout is the template's). Also what the person pays to export it (export.priceUsd, export.tellThePerson).

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The Pro template's id (from list_pro_templates)."
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
🟡create_pro_draft(template, ops)

Make a draft of a Pro template with the brand's texts, images and colors: set (a layer's text, its image src and fit, its colors, or hidden), recolor (one palette color to another everywhere) and video (its background). The layout, timing and motion stay the template's: an operation that changes them is refused, and the person can change them in the editor. All are checked together: one that fails applies nothing and the answer lists every issue. Returns draftId, editToken (shown once), editorUrl (opens the Pro editor, where the person watches it, can change anything and exports it), warnings and export (its price, to tell the person).

入力スキーマ

{
  "type": "object",
  "properties": {
    "template": {
      "type": "string",
      "description": "The Pro template's id."
    },
    "ops": {
      "type": "array",
      "items": {
        "description": "set (a layer's text, props.text/src/fit/color/fill/stroke/background, or hidden), recolor (from one color to another everywhere), or video (its background). Where things sit, when they show and how they move are the template's.",
        "anyOf": [
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "set"
                ]
              },
              "layer": {
                "type": "string",
                "description": "A layer id (get_pro_template lists them, with what each is for)."
              },
              "format": {
                "type": "string",
                "description": "\"all\" (default) or one format id."
              },
              "text": {
                "type": "string",
                "description": "A text layer's words."
              },
              "props": {
                "type": "object",
                "properties": {
                  "text": {
                    "type": "string"
                  },
                  "src": {
                    "type": "string",
                    "description": "An image layer: the template's own asset or an upload_pro_image src."
                  },
                  "fit": {
                    "type": "string",
                    "enum": [
                      "cover",
                      "contain"
                    ]
                  },
                  "color": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "fill": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "stroke": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "background": {
                    "type": "string",
                    "description": "#RRGGBB"
                  }
                },
                "additionalProperties": false
              },
              "hidden": {
                "type": "boolean",
                "description": "true hides the layer (a part the brand doesn't need)."
              }
            },
            "required": [
              "op",
              "layer"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "recolor"
                ]
              },
              "from": {
                "type": "string",
                "description": "#RRGGBB"
              },
              "to": {
                "type": "string",
                "description": "#RRGGBB"
              }
            },
            "required": [
              "op",
              "from",
              "to"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "video"
                ]
              },
              "background": {
                "type": "string",
                "description": "#RRGGBB"
              }
            },
            "required": [
              "op",
              "background"
            ],
            "additionalProperties": false
          }
        ]
      },
      "description": "Operations, for example [{\"op\":\"set\",\"layer\":\"title\",\"text\":\"Plan less.\\nShip more.\"},{\"op\":\"recolor\",\"from\":\"#FE7B02\",\"to\":\"#5E6AD2\"}]. The guide (resource opusmotion://guide/pro) says how to make it good."
    }
  },
  "required": [
    "template"
  ],
  "additionalProperties": false
}
🟡edit_pro_draft(draftId, editToken, ops, revision)

Change a Pro draft's texts, images and colors (the same operations as create_pro_draft: set, recolor, video's background). Send revision to be sure nobody changed it since you read it (a stale one is refused). Returns the new revision and warnings.

入力スキーマ

{
  "type": "object",
  "properties": {
    "draftId": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "editToken": {
      "type": "string",
      "description": "The editToken the draft came with."
    },
    "ops": {
      "type": "array",
      "items": {
        "description": "set (a layer's text, props.text/src/fit/color/fill/stroke/background, or hidden), recolor (from one color to another everywhere), or video (its background). Where things sit, when they show and how they move are the template's.",
        "anyOf": [
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "set"
                ]
              },
              "layer": {
                "type": "string",
                "description": "A layer id (get_pro_template lists them, with what each is for)."
              },
              "format": {
                "type": "string",
                "description": "\"all\" (default) or one format id."
              },
              "text": {
                "type": "string",
                "description": "A text layer's words."
              },
              "props": {
                "type": "object",
                "properties": {
                  "text": {
                    "type": "string"
                  },
                  "src": {
                    "type": "string",
                    "description": "An image layer: the template's own asset or an upload_pro_image src."
                  },
                  "fit": {
                    "type": "string",
                    "enum": [
                      "cover",
                      "contain"
                    ]
                  },
                  "color": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "fill": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "stroke": {
                    "type": "string",
                    "description": "#RRGGBB"
                  },
                  "background": {
                    "type": "string",
                    "description": "#RRGGBB"
                  }
                },
                "additionalProperties": false
              },
              "hidden": {
                "type": "boolean",
                "description": "true hides the layer (a part the brand doesn't need)."
              }
            },
            "required": [
              "op",
              "layer"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "recolor"
                ]
              },
              "from": {
                "type": "string",
                "description": "#RRGGBB"
              },
              "to": {
                "type": "string",
                "description": "#RRGGBB"
              }
            },
            "required": [
              "op",
              "from",
              "to"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "op": {
                "type": "string",
                "enum": [
                  "video"
                ]
              },
              "background": {
                "type": "string",
                "description": "#RRGGBB"
              }
            },
            "required": [
              "op",
              "background"
            ],
            "additionalProperties": false
          }
        ]
      }
    },
    "revision": {
      "type": "integer",
      "description": "The revision you read."
    }
  },
  "required": [
    "draftId",
    "editToken",
    "ops"
  ],
  "additionalProperties": false
}
🟢get_pro_draft(draftId)

A Pro draft as it is now: its revision, every layer (what differs per format) and both formats' docs.

入力スキーマ

{
  "type": "object",
  "properties": {
    "draftId": {
      "type": "string",
      "description": "The draft's id (draftId)."
    }
  },
  "required": [
    "draftId"
  ],
  "additionalProperties": false
}
🟡upload_pro_image(draftId, editToken, data, contentType, file)

Store an image (PNG, JPEG, WebP or SVG) in a Pro draft and get its src for an image layer ({"op":"set","layer":"hero","props":{"src":"<src>"}}). Send small images as data (base64 or a data URL, up to 3 MB); for a file on disk, leave data out and the answer is the curl command that uploads it.

入力スキーマ

{
  "type": "object",
  "properties": {
    "draftId": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "editToken": {
      "type": "string",
      "description": "The editToken the draft came with."
    },
    "data": {
      "type": "string",
      "description": "The image as base64 or a data URL."
    },
    "contentType": {
      "type": "string",
      "enum": [
        "image/png",
        "image/jpeg",
        "image/webp",
        "image/svg+xml"
      ],
      "description": "Its type, when data is plain base64."
    },
    "file": {
      "type": "string",
      "description": "Without data: the file's path, to write it into the curl command."
    }
  },
  "required": [
    "draftId",
    "editToken"
  ],
  "additionalProperties": false
}
🟡look_at_pro_draft(draftId, editToken, formats, times, sheet, ...)

Render stills (or a contact sheet) of a Pro draft as it is now, in each format, and return them as images inside the answer, with a layout check of those moments by layer (text hidden under another layer, past its box, low contrast). Look before you hand the link over; change and look again.

入力スキーマ

{
  "type": "object",
  "properties": {
    "draftId": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "editToken": {
      "type": "string",
      "description": "The editToken the draft came with."
    },
    "formats": {
      "description": "Which formats: \"all\", or a list such as [\"9:16\"]. Pro drafts default to all; catalog drafts to the draft's own.",
      "anyOf": [
        {
          "type": "string",
          "enum": [
            "all"
          ]
        },
        {
          "type": "array",
          "items": {
            "type": "string",
            "description": "A format id the template offers, for example 16:9, 9:16 or 1:1."
          }
        }
      ]
    },
    "times": {
      "type": "array",
      "items": {
        "type": "number",
        "minimum": 0
      },
      "description": "Moments in seconds. Stills: up to 6 (default: the middle of each act or scene and the closing hold). Sheet: up to 9 (default: 9 evenly spread)."
    },
    "sheet": {
      "type": "boolean",
      "description": "One contact sheet per format (the moments side by side, labelled) instead of separate stills: the motion at a glance for few tokens."
    },
    "wait": {
      "type": "number",
      "minimum": 0,
      "maximum": 55,
      "description": "Seconds to wait for the render before answering (default 50). If it isn't ready, the answer has a jobId for get_frames."
    },
    "inline": {
      "type": "boolean",
      "description": "Return the images inside the answer (default true). false returns their URLs only."
    }
  },
  "required": [
    "draftId",
    "editToken"
  ],
  "additionalProperties": false
}
🟢look_at_draft(id, editToken, formats, times, sheet, ...)

Render stills (or a contact sheet) of a catalog draft with its texts, images and adjustments, the way its export will, and return them as images inside the answer with a layout check of those moments.

入力スキーマ

{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "editToken": {
      "type": "string",
      "description": "The editToken the draft came with."
    },
    "formats": {
      "description": "Which formats: \"all\", or a list such as [\"9:16\"]. Pro drafts default to all; catalog drafts to the draft's own.",
      "anyOf": [
        {
          "type": "string",
          "enum": [
            "all"
          ]
        },
        {
          "type": "array",
          "items": {
            "type": "string",
            "description": "A format id the template offers, for example 16:9, 9:16 or 1:1."
          }
        }
      ]
    },
    "times": {
      "type": "array",
      "items": {
        "type": "number",
        "minimum": 0
      },
      "description": "Moments in seconds. Stills: up to 6 (default: the middle of each act or scene and the closing hold). Sheet: up to 9 (default: 9 evenly spread)."
    },
    "sheet": {
      "type": "boolean",
      "description": "One contact sheet per format (the moments side by side, labelled) instead of separate stills: the motion at a glance for few tokens."
    },
    "wait": {
      "type": "number",
      "minimum": 0,
      "maximum": 55,
      "description": "Seconds to wait for the render before answering (default 50). If it isn't ready, the answer has a jobId for get_frames."
    },
    "inline": {
      "type": "boolean",
      "description": "Return the images inside the answer (default true). false returns their URLs only."
    }
  },
  "required": [
    "id",
    "editToken"
  ],
  "additionalProperties": false
}
🟢get_frames(draftId, jobId, wait, inline)

The images of a look that wasn't ready when it was asked (look_at_draft or look_at_pro_draft answered a jobId).

入力スキーマ

{
  "type": "object",
  "properties": {
    "draftId": {
      "type": "string",
      "description": "The draft's id (draftId)."
    },
    "jobId": {
      "type": "string",
      "description": "The jobId the look answered."
    },
    "wait": {
      "type": "number",
      "minimum": 0,
      "maximum": 55,
      "description": "Seconds to wait for the render before answering (default 50). If it isn't ready, the answer has a jobId for get_frames."
    },
    "inline": {
      "type": "boolean",
      "description": "Return the images inside the answer (default true). false returns their URLs only."
    }
  },
  "required": [
    "draftId",
    "jobId"
  ],
  "additionalProperties": false
}

コミュニティ

このサーバーを評価する

エビデンス

最近の観測

検証済みバージョンは記録されていませんツール 17 件