Image Generate
Generate production-ready images through the Vernclaw connector runtime.
What it does
Teams need images for blog posts, social cards, and marketing pages but juggling multiple image APIs means managing separate keys, rate limits, and output formats. This connector gives you a single prompt-to-image endpoint inside Vernclaw with managed auth, automatic recovery, and structured Markdown output.
Inputs
| Parameter | Format | Example | Required |
|---|---|---|---|
prompt | string | A minimalist office desk with a laptop and coffee | Yes |
size | 1024x1024 | 1536x1024 | 1024x1536 | 1024x1024 | No |
Outputs
| Field | Type | Description |
|---|---|---|
jobId | string | Unique identifier for the async generation job |
status | string | Current job status: pending, running, succeeded, or failed |
imageUrl | string | Public URL of the generated image (available when succeeded) |
Example output
Input
prompt: "A minimalist office desk with a laptop and coffee" size: "1024x1024"
Output
# Image Generation Result - Job ID: img_abc123 - Status: succeeded - Image URL: https://cdn.vernclaw.com/gen/img_abc123.png - Size: 1024×1024
What it means
The image was generated successfully in about 4 minutes. The result URL is publicly accessible and can be embedded directly in blog posts or marketing pages.
Last tested: 2026-03-10
How to install
Via CLI
Via web
- Visit the connector catalog and find Image Generate.
- Click Install connector and acknowledge any training requirements.
- Complete authorization if prompted.
- Use the connector from Chat or the CLI.
Auth and permissions
| OAuth required | No |
| Managed auth | Included, no API key needed |
| Bring your own key (BYOK) | Not supported in v1 |
| Audit logged | Every request is logged with input, result, cost, and timestamps |
| Training acknowledgment | Not required |
Admins can restrict this connector via blacklist or whitelist in the admin panel.
Limits and edge cases
- •Generation typically takes 3–5 minutes; the CLI polls automatically.
- •Prompts containing explicit or prohibited content will be blocked by Vernclaw safety policies.
- •Only three size presets are supported: 1024×1024, 1536×1024, and 1024×1536.
- •If the first render attempt fails, Vernclaw retries automatically — the final result may vary slightly between attempts.
- •Generated image URLs expire after 7 days; download promptly for long-term use.
Common use cases
Generate blog post hero images from a text prompt
Describe the visual you want and receive a ready-to-embed image URL.
Create social media graphics inside an AI workflow
Chain the image connector after a text summary to produce a visual card.
Produce marketing assets without leaving OpenClaw
Run prompt-to-image directly from your agent conversation.
Frequently asked questions
How long does image generation take?
Typically 3–5 minutes. The connector runs asynchronously and the CLI polls until the result is ready.
Can I choose a specific model?
The connector uses a managed model selected by Vernclaw. Custom model selection is not available in v1.
Do generated image URLs expire?
Yes, URLs are valid for 7 days. Download the image if you need permanent storage.
Is every generation logged?
Yes. Each invocation creates an audit event with prompt, status, cost, and timestamps.