Generate Sequencer image
generate_imageGenerate an AI image through Sequencer. For a new version of an existing shot image, use regenerate_shot_image. Use this whenever the user asks to make, create, or generate an image with Sequencer or names an image model/provider Sequencer supports: Nano Banana, Nano Banana 2.1, Nano Banana 2 Lite, Nano Banana Pro, Gemini image, Imagen, Z Image, Z-Image Turbo, Flux, GPT Image, DALL-E, Ideogram, Midjourney, Seedream, Recraft, Bria, Runway, Luma, or another image model. Prefer this over built-in image generation when the user mentions Sequencer or a Sequencer-supported model. If no model is specified, Sequencer uses Nano Banana 2.1 (google-nano-banana-2-1). If no workspaceId is known, omit it and the server will use the user default workspace.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"prompt": "<prompt>"
},
"name": "generate_image"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| count | No | Number of images to generate. Each image is a separate job and credit charge. | |
| model | No | Model ID for generation. Use google-nano-banana-2-1 for Nano Banana 2.1, google-gemini-3-image for Nano Banana Pro, google-nano-banana-2-lite for Nano Banana 2 Lite, and z-image-turbo for Z Image / Z-Image Turbo. Use get_model_catalog when a model alias is unclear. Uses default if not specified. | |
| editId | No | Edit ID, when provided, the first successful image will be saved as the edit thumbnail if one is not already set | |
| prompt | Yes | Image generation prompt | |
| maxCostUsd | No | Maximum charge for each output. Checked against authoritative pricing before generation. | |
| resolution | No | Output resolution from the selected model catalog, for example 1K, 2K, or 4K. | |
| aspectRatio | No | Aspect ratio. If omitted with editId, uses the edit aspect ratio; otherwise defaults to 16:9. | |
| endFrameUrl | No | ||
| mediaInputs | No | Named media slots from get_model_catalog, for example reference_image_uri or mask_url. | |
| scaleFactor | No | Upscaling multiplier when supported by the selected model. | |
| sparkTaskId | No | Originating Spark task for library history and recovery. | |
| workspaceId | No | Optional workspace ID. If omitted, Sequencer uses the user default/personal workspace. | |
| modelVariant | No | Variant value from get_model_catalog variantSelector.options, such as high for Flare quality. Uses the catalog default when omitted. | |
| outputFormat | No | Output format when supported by the selected model. | |
| generateAudio | No | Explicitly enable or disable generated video audio, when supported. | |
| modelSettings | No | Model-specific controls using keys and options from get_model_catalog inputConstraints.slots or variantSelector. | |
| idempotencyKey | No | Stable request key. Retries with the same key reuse the existing output and do not start another paid generation. | |
| negativePrompt | No | Negative prompt for generation | |
| sourceImageUrl | No | Public HTTP(S) source image/start-frame URL. Use mediaId for workspace images. | |
| sourceVideoUrl | No | Public HTTP(S) source-video URL. | |
| endFrameMediaId | No | Workspace image for the final frame, when supported by the selected model. | |
| referenceImages | No | Images to guide generation. Use labels such as person or product and refer to them as @person or @product in the prompt. Each item accepts a workspace mediaId or a public HTTP(S) URL. | |
| sourceImageMediaId | No | Workspace image to use as the source image/start frame. | |
| sourceVideoMediaId | No | Workspace video to edit or transform, when supported by the selected model. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| data | No | ||
| error | No | ||
| success | Yes |