2.6k
GitHub Stars
3
Bundled Files
5 months ago
Catalog Refreshed
6 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill ima-image-ai- _meta.json653 B
- requirements.txt111 B
- SKILL.md62.7 KB
Overview
This skill integrates with the IMA Open API to generate images from text prompts or existing images. It picks the newest and most popular model by default (SeeDream 4.5) while offering budget and premium options. Use it when you need clear cost/time estimates, progress updates, and delivered images rendered directly in chat. A valid ima_* API key is required.
How this skill works
The agent sends a prompt or source image to IMA Studio, starts a background generation task, and polls for status while pushing user-facing updates. Defaults: text_to_image and image_to_image use SeeDream 4.5 (doubao-seedream-4.5, 5 pts). Other available models: Nano Banana2 (4 pts, budget), Nano Banana Pro (10–18 pts, premium with resolution choices), and Midjourney (8–10 pts, artist-level). When done, the image is posted directly with a caption and an original-link for easy sharing.
When to use it
- Generate high-quality balanced images quickly using the default SeeDream 4.5.
- Choose Nano Banana2 when you want the fastest, lowest-cost runs.
- Choose Nano Banana Pro for premium quality and higher-resolution outputs (1K/2K/4K).
- Use Midjourney for artist-style, aesthetic-first results.
- Perform image-to-image editing by supplying a source image and prompt.
Best practices
- Always recommend the newest, most popular model by default (SeeDream 4.5) unless user requests budget/premium.
- Send an immediate short acknowledgement reply first, then push start/progress/result updates to avoid ordering issues.
- Provide clear model, estimated time, and credit cost in the start notification.
- Send periodic progress updates capped at 95% and display actual runtime on success.
- Translate technical errors into natural-language suggestions and offer 1–2 alternative models.
Example use cases
- Text-to-image: Create a 16:9 landscape with SeeDream 4.5 (default) and receive the rendered image in chat.
- Budget run: Quick social media asset using Nano Banana2 to minimize credits and latency.
- Premium output: High-resolution poster with Nano Banana Pro at 2K or 4K.
- Artistic render: Use Midjourney for stylized, artist-level compositions.
- Image-to-image: Improve or re-style an existing photo using SeeDream 4.5 (default i2i attribute id 1611).
FAQ
Default selection follows the newest and most popular model: SeeDream 4.5 for both text-to-image and image-to-image.
Can I pick a cheaper model?
Yes — Nano Banana2 is the budget option (4 pts). Specify it explicitly if you prefer lower cost and faster generation.
What if generation fails?
You’ll get a natural-language failure message with suggested alternative models (typically SeeDream 4.5 and Nano Banana2).