Capabilities · Last updated September 23, 2026
Image Generation for AI agents
Generate an image from a prompt or revise an existing asset through the AnyCap HTTP API or CLI. Use the JSON API in your own application, or let Claude Code, Cursor, Codex, and other agents call the CLI without building a separate integration for each model.
One real generation, then an edit from the saved image
Production CLI trial, September 22, 2026. We ran Seedream 5 with AnyCap CLI 0.6.2 against the production API. The first command generated an olive desk lamp; the second used that file as an image reference and asked for a yellow shade. Both requests succeeded and returned 2848 × 1600 PNG files. The previews below are resized JPEG copies of those outputs.
The shade changed, but the edit also introduced a small mark near the base. In this one run, the CLI reported 2 credits for each request; that is an observed charge, not a price guarantee. This is evidence for the CLI workflow, not an executed result for the HTTP request example on this page.
See the exact commands, review steps, and limits →Which model should you start with?
Choose text-to-image for a new asset or image-to-image when you have a reference to revise. The Seedream 5 generation and edit below show one tested workflow; compare other models against your own brief before choosing one for repeated use.
- Agents can create first-pass visuals, revise source images, and keep asset delivery in one AnyCap workflow.
- Text-to-image and image-to-image modes stay behind one AnyCap command surface.
- Model choice stays explicit, from first-pass generation to revision-heavy image editing.
Checked against the live catalog · Production catalog verified October 21, 2026
Thirteen image models, checked from the production CLI.
We rechecked the production catalog with AnyCap CLI 0.6.5 on October 21, 2026: 13 active image models. Twelve declare text-to-image and image-to-image; Seedream 5 Pro Multi-Reference declares image-to-image only. This verifies availability and declared modes, not generation quality across models.
- What we found
- 13 active models
- Operations
- generate
- Declared modes
- 12 both · 1 image-to-image only
- CLI version
- 0.6.5
anycap image modelsModels grouped under image generation
Choose a model by the input mode and request fields your job needs. The catalog is not a quality ranking; compare outputs on your own brief and check the live schema before production.
- GPT Image 2text-to-image, image-to-image
- Nano Banana Protext-to-image, image-to-image
- Nano Banana 2text-to-image, image-to-image
- Seedream 5text-to-image, image-to-image
- Seedream 5 Protext-to-image, image-to-image
- Seedream 5 Pro Multi-Referenceimage-to-image
- FLUX.1 Kontext Maxtext-to-image, image-to-image
- Qwen Imagetext-to-image, image-to-image
- Nano Banana 2 Litetext-to-image, image-to-image
- Seedream 4.5text-to-image, image-to-image
- GPT Image 2.5 Flaretext-to-image, image-to-image
- GPT Image 2.5 Sunbursttext-to-image, image-to-image
- Nano Banana 2.1text-to-image, image-to-image
Models change. AnyCap keeps discovery, supported modes, authentication, execution, and output delivery behind one agent-facing interface, so an agent can pick a valid option today and switch when the catalog changes.
HTTP API and CLI usage
HTTP text-to-image request (requires an API key)
curl -X POST https://api.anycap.ai/v1/image/generate \
-H "Authorization: Bearer $ANYCAP_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"seedream-5","mode":"text-to-image","prompt":"a minimalist product hero image on a cream background"}'CLI text-to-image
anycap image generate --prompt "a minimalist product hero image on a cream background" --model seedream-5 -o hero.pngImage-to-image editing
anycap image generate --prompt "turn this into a warm editorial product shot" --model nano-banana-pro --mode image-to-image --param images=./source.png -o variation.pngDiscover models
anycap image modelsHow image generation fits an AnyCap workflow
Brief
The agent turns a product, creator, or design request into a prompt and chooses whether the job starts from text or an existing image.
Generate
AnyCap runs the selected image model with the right mode, model ID, prompt, and output file.
Iterate
The result can move into review, editing, Drive delivery, Page publishing, or a follow-up image-to-video workflow.
When agents and creators need image generation
Product mockups
Generate polished visuals for launch pages, changelogs, and internal demos.
Creative iteration
Run text-to-image and image editing loops without leaving the agent workflow.
Creators and marketers
Create illustrations, thumbnails, social posts, and marketing assets through one repeatable command surface.
Everyday edits
Turn briefs, screenshots, and references into first-pass visual directions, background swaps, and simple photo edits.
FAQ
What does AnyCap image generation let agents do?
It gives agents one command surface for text-to-image and image-to-image workflows. That means the same CLI can handle first-pass generation, creative iteration, and image editing without separate provider integrations.
Which image models are available through AnyCap today?
The September 23, 2026 CLI catalog returned ten active image models, including Seedream 5, Seedream 5 Pro, Nano Banana Pro, Nano Banana 2, GPT Image 2, FLUX.1 Kontext Max, and Qwen Image. Nine list both text-to-image and image-to-image; Seedream 5 Pro Multi-Reference lists image-to-image only. Check the live catalog and the selected mode schema before sending a request.
Why does this page mention image editing as well as image generation?
Market language often splits text-to-image, image editing, and image generation. AnyCap groups those workflows under one image generation capability because agents frequently need both creation and revision in the same loop.
Is this page about an image generation API or a CLI?
Both. The HTTP endpoint is POST /v1/image/generate with a Bearer API key and JSON fields such as model, mode, and prompt. The CLI provides the same capability inside coding-agent workflows. This request example describes the interface; it is not a recorded generation result.
How do I know which image options a model accepts?
Check the selected model's mode schema before adding fields such as aspect_ratio, resolution, or reference images. Supported options and limits vary by model and mode; the CLI command is anycap image models <model-id> schema --mode <mode>.
Is this only for developers?
No. The same capability supports creators, marketers, operators, and everyday users who need product visuals, social content, thumbnails, or quick photo edits. The agent workflow is just one of the ways to reach it.
Let your agent create the visual.
Use AnyCap when image generation, editing, model selection, and asset delivery should stay inside the same agent workflow.
CLI version and updating · 0.6.5
Current release checked October 21, 2026: 0.6.5. Use the update command for your installation method, then verify the version and login in the same environment that runs your agent.
Direct binary installation
anycap updatenpm installation
npm install -g @anycap/cli@latestanycap --version
anycap statusVersion 0.6.5 retries transient downloads, recovers an expired video link by querying the same task, and saves multiple image outputs to numbered files. After a video download failure, resume the existing task instead of generating again.
anycap video tasks wait <task-id> -o result.mp4Dated examples retain the CLI version, output and charge from the original run. Updating the CLI does not revalidate an old result. Check the live model schema and current pricing before a new billed task.