weryai-image-generator（维雅图像生成器）

WeryAI Image Generator

Generate WeryAI images with the official base skill for text-to-image and image-to-image workflows. In agent environments, default to an image-first flow: submit the task, then poll status until final images are ready or the maximum timeout of 5 minutes (300 seconds) is reached. Do not run unbounded polling loops. Treat taskId or batchId as tracking data, not the endpoint, and only surface them as the primary output when the user explicitly wants IDs first or timeout is reached before completion. Use wait-image.js only when the user explicitly asks for one-shot submit-and-wait behavior.

Example Prompts

• Generate this image and keep checking until you can show me the final image, or stop when the 5-minute timeout is reached.
• Turn this reference image into a cinematic poster, and check the result only if I ask for status.
• Restyle this image with WeryAI image-to-image and show me the final image when it is ready.
• Check which WeryAI image models support 9:16 and 4 output images before generating.
• Check my WeryAI image generation task status and tell me whether the images are ready yet.

Quick Summary

• Main jobs: text-to-image, image-to-image, image from prompt, restyle this image, task status, final image delivery
• Default model: WeryAI Image 2.0 (WERYAI_IMAGE_2_0)
• Default parameters: image_number=1, aspect_ratio=9:16
• Main trust signals: dry-run support, model capability lookup, paid-run warning, media-source validation with auto upload for local references

Authentication and first-time setup

Before the first real generation run:

1. Create a WeryAI account.
2. Open the API key page at https://www.weryai.com/api/keys.
3. Create a new API key and copy the secret value.
4. Add it to the required environment variable WERYAI_API_KEY.
5. Make sure the WeryAI account has available balance or credits before paid generation.

OpenClaw-friendly setup

• This skill already declares WERYAI_API_KEY in metadata.openclaw.requires.env and primaryEnv.
• After installation, if the installer or runtime asks for required environment variables, paste the key into WERYAI_API_KEY.
• If you are configuring the runtime manually, export it before running commands:

export WERYAI_API_KEY="your_api_key_here"

Quick verification

Use one safe check before the first paid run:

node scripts/models-image.js --mode text_to_image
node scripts/wait-image.js --json '{"prompt":"A glowing paper lantern in mist","aspect_ratio":"9:16"}' --dry-run

• models-image.js confirms that the key is configured and the models endpoint is reachable.
• --dry-run confirms the request shape locally without spending credits.
• Real wait or submit-* commands still require available WeryAI balance.

Prerequisites

• WERYAI_API_KEY must be set before paid runs.
• Node.js >=18 is required because the runtime uses built-in fetch.
• For image-to-image, reference images may be http/https URLs or local/file sources. Local/non-http(s) sources are uploaded first via /v1/generation/upload-file.
• Real submit and wait runs consume WeryAI credits.

Security And API Hosts

• Keep WERYAI_API_KEY secret and never write it into the repository.
• This skill supports directly passing local file paths. If a local file path is provided, the runtime will automatically upload the local file to the WeryAI server for processing.
• Optional overrides WERYAI_BASE_URL and WERYAI_MODELS_BASE_URL default to https://api.weryai.com and https://api-growth-agent.weryai.com. Only override them with trusted hosts.
• Review scripts/ before production use if you need higher assurance.

Supported Intents

• Text brief -> generate one or more images from scratch.
• One reference image -> restyle or transform that image.
• Existing task -> check image generation status instead of creating a new paid job.
• Parameter or model question -> inspect models first, then submit.

Default Configuration

Unless the user explicitly changes them, prefer:

• model: WERYAI_IMAGE_2_0 (WERYAI_IMAGE_2_0)
• image_number: 1
• aspect_ratio: 9:16

Always allow the user to override model, image_number, aspect_ratio, and resolution. When the user asks for unsupported settings, run models-image.js and keep only values supported by the chosen model.

Model Switching And Parameter Guidance

Guide the user progressively instead of explaining every parameter up front.

• If the user only wants an image, proceed with the default WeryAI Image 2.0 (WERYAI_IMAGE_2_0) configuration.
• If the user asks for a different model, better quality, more images, another aspect ratio, or a specific resolution, switch into parameter-confirmation mode.
• If the user already knows the exact parameter they want, apply it directly and only validate model support when needed.
• If the user sounds unsure, translate their creative request into the closest supported parameters rather than asking them to choose raw API fields.

Recommended guidance pattern

Use short operator-style guidance like this:

• General help: When the user asks "how to use this skill", DO NOT paste raw shell commands. Instead, explain the capabilities in natural language and give 2-3 prompt examples.
• Default run:

I can start with the default setup: WeryAI Image 2.0, 1 image, 9:16. If you want, I can also switch the model or adjust the image count, aspect ratio, or resolution before submission.

• Model switching:

If you want another model, tell me whether you care more about realism, stylization, reference-image fidelity, higher resolution, or output count, and I will check the supported models first.

• Parameter changes:

I can map your request into image settings. For example: vertical cover -> 9:16, square post -> 1:1, more options -> image_number 4, clearer output -> the highest supported resolution for that model.

• Safety before paid runs:

Before I submit a paid task, I will show the final model, parameters, and prompt so you can confirm them.

When to ask follow-up questions

Ask only for the smallest missing detail needed to submit safely.

• Ask about aspect_ratio when the user implies platform intent such as poster, cover image, wallpaper, square post, or vertical social card.
• Ask about image_number when the user wants multiple options or variations.
• Ask about resolution only when the user explicitly asks for higher quality or the target model supports multiple resolution tiers.
• Ask about model choice only when the user explicitly wants a different model or when capability support is uncertain.
• Do not ask every parameter question if the default configuration already fits the request.

Map user language to parameters

Use these common mappings:

• vertical, poster, cover image, mobile cover, social card -> aspect_ratio: 9:16
• square, avatar, icon, social post -> aspect_ratio: 1:1
• landscape, wide banner, wallpaper, hero image -> choose a supported wide aspect ratio such as 16:9
• give me a few options, multiple versions, more variations -> increase image_number
• make it clearer, higher quality, higher resolution -> use the highest supported resolution for the chosen model
• use another model, check supported models, switch model -> run models-image.js before submission

Model lookup workflow

When the user asks to change the model or requests parameters that may be unsupported:

1. Identify the intended mode: text_to_image or image_to_image.
2. Run the matching models-image.js command first.
3. Keep only supported values for image_number, aspect_ratio, and resolution.
4. If multiple models fit, recommend one concise default instead of dumping raw metadata unless the user asked for a comparison.
5. If support is unclear, say so explicitly and use --dry-run or a safe model query before the paid call.

Confirmation block before submission

Before a paid run, show a concise confirmation block with the final payload choices.

Ready to generate

- mode: `image-to-image`
- model: `WERYAI_IMAGE_2_0`
- image_number: `1`
- aspect_ratio: `9:16`
- resolution: `default`
- image: `https://example.com/input.png`
- prompt: `Restyle this portrait into a cinematic editorial image, preserve facial identity, refined lighting, clean composition, premium color grading, polished final output.`

Wait for confirmation or requested edits before running a paid submission.

Intent Routing

Use image-first submit plus bounded status polling as the default path in agent environments.

• If the user provides only prompt, route to text-to-image.
• If the user provides image or images, route to image-to-image.
• If the user already has taskId or batchId, use status-image.js instead of creating a new task.
• If the user asks about supported models or parameters, run models-image.js before any paid submission.
• After submit-*, run status-image.js polling by default until final images are ready or the 5-minute timeout is reached.
• Use wait-image.js only when the user explicitly asks for a blocking one-shot result.
• Do not perform unbounded status polling; enforce the 5-minute timeout ceiling for default polling.
• If the task is still processing, keep the user informed, but do not treat the task ID as the final deliverable.

Preferred Commands

# Default async submit
node scripts/submit-text-image.js --json '{"prompt":"A refined editorial portrait"}'

# Poll an existing task
node scripts/status-image.js --task-id <task-id>

Workflow

1. Identify the user's intent: text-to-image, image-to-image, status lookup, or model lookup.
2. Collect the prompt and, if needed, one or more public https reference image URLs.
3. Apply defaults: WeryAI Image 2.0 (WERYAI_IMAGE_2_0), image_number=1, aspect_ratio=9:16, unless the user asks otherwise.
4. If the user wants a custom model or non-default parameters, run models-image.js first when support is uncertain.
5. Use --dry-run when you need to preview the final payload before a paid submission.
6. Default to image-first execution:

- Stage 1: submit-*. - Stage 2: run status-image.js polling until images are ready or the 5-minute timeout is reached. - Stage 3: if timeout is reached, return taskId/batchId with a timeout note and a follow-up status command.

1. Do not treat status lookup as an unbounded loop; always enforce timeout ceilings.
2. Use status-image.js to re-check an existing task or batch safely.

Input Rules

• prompt is required for both text-to-image and image-to-image requests.
• For image-to-image, either image or images is accepted; both are normalized to the API images array.
• Every reference image source can be http/https URL or local/file source (local/file sources are uploaded first).
• Prefer model-supported values for aspect_ratio, image_number, and resolution.
• Do not invent undocumented request fields.

Output

All commands print JSON to stdout. Successful result objects can include:

• taskId, taskIds, batchId
• taskStatus
• images
• balance
• errorCode, errorMessage

User-facing delivery requirement:

• If image URLs are available, return at least one user-visible image link (for example [Image](https://...)) or inline image rendering when supported. If multiple images are generated, render all of them using markdown image syntax consecutively.
• Alongside image output, include key generation parameters when available: model, aspect_ratio, image_number, and resolution.
• Do not use taskId / batchId as the sole deliverable unless the user explicitly requested IDs first.
• If timeout is reached before completion, return the taskId to the user and ask if they want you to check the status again. Do NOT show the raw node status command to the user; use it internally.

See references/error-codes.md for common failure classes and recovery hints.

Definition Of Done

The task is done when:

• local validation passes without CLI-side errors,
• the user can see at least one generated image or a usable download URL,
• or wait-image.js reaches a terminal result with at least one image URL,
• or status-image.js polling reaches timeout and returns a clear in-progress status with task tracking data,
• and the output makes it explicit whether image URLs are present.
• When only task tracking data is available due to timeout, the reply must include the timeout reason and an explicit follow-up status command.

Constraints

• Do not assume every image model supports every optional field.
• Do not use local file paths for reference images.
• Do not re-run submit or wait casually because each run can create a new paid task.
• Do not default to wait-image.js in agent environments for generation tasks that may block the session.
• Do not broaden this skill into general image editing outside the documented WeryAI API surface.

Re-run Behavior

• submit-text-image.js and submit-image-to-image.js are not idempotent; re-running them can create new paid tasks.
• wait-image.js is not idempotent for the same reason: it submits first, then polls until a terminal result or timeout.
• status-image.js, models-image.js, and balance-image.js are safe to re-run.

References

• Model metadata and parameter rules: references/api-models.md
• Error handling guidance: references/error-codes.md