GET
/v1/schemas/byteplus/image/images/generations
Accept:
| field | recorded source |
|---|---|
| schema | Source linked above. generated Source tracee3379046f28d4628702fa094c1ce8678f3e103e6579a5dad417f19e5c90f67e3{
"fetchedAt": 1787893860,
"extractorVersion": "1",
"verifiedAt": null
} |
Request body
16 properties · 2 required · 8 $defs| property | type | description | constraints |
|---|---|---|---|
| model* | string | Seedream model id (e.g. seedream-4-0-250828) or a preconfigured endpoint id. | |
| prompt* | string | Instruction text (docs give the limit as 600 English words). | |
| image | any | Input images for image-conditioned generation (editing, reference-guided generation, multi-reference composition). Each entry is a publicly reachable URL or a data:image/<format>;base64,<data> URI — <format> must be lowercase. Up to 14 references (10 on dola-s… | |
| response_format | string | How generated images come back. 'url' links expire 24 hours after generation. Server default 'url'. | |
| seed | integer | ||
| guidance_scale | number | How strictly the result follows the prompt; higher values adhere more closely at some cost to diversity. | |
| size | string | Output size: a shorthand token ('1K', '2K', '4K') or explicit pixels ('2048x2048') — never a mix. Server default 2048x2048. | |
| watermark | boolean | Stamp an 'AI generated' watermark in the bottom-right corner. DEFAULTS TO TRUE — pass false to opt out. | |
| optimize_prompt | boolean | ||
| optimize_prompt_options | OptimizePromptOptions | ||
| sequential_image_generation | string | Group-image mode: 'auto' lets the model return a set of related images (bounded by sequential_image_generation_options.max_images — an upper bound, not a count); 'disabled' (server default) returns exactly one. | |
| sequential_image_generation_options | SequentialImageGenerationOptions | ||
| tools | ContentGenerationTool[] | Provider tools to run during generation; the only documented member is web search, whose usage comes back under usage.tool_usage.web_search. | |
| output_format | enum | Generated file format. Seedream 5.0 family only; default 'jpeg'. | jpeg | png |
| layer_decomposition | boolean | ||
| stream | boolean | Server-sent-events mode, emitting each image as it finishes. Absent from the Go SDK request struct, which models the unary call only. |
Types
8 $defsContentGenerationToolobject
| type* | enum | web_search |
ContentGenerationToolTypeenum
web_search
OptimizePromptModestring
OptimizePromptOptionsobject
| thinking | string | ||
| mode | string |
OptimizePromptThinkingstring
OutputFormatenum
jpeg | png
SequentialImageGenerationstring
SequentialImageGenerationOptionsobject
| max_images | integer |
Validate a payload
before you spend tokensshell
$ curl -X POST https://modelschemas.com/v1/validate \
-d '{"provider":"byteplus","endpointId":"images/generations","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}