Text to Image API

Generate images from text with one API call

Flexura's Image Generation API turns text prompts into high-quality images using NVIDIA FLUX. Async job processing means your agent never blocks — submit a prompt, poll for results, and get a download URL.

NVIDIA FLUX powered
Async job queue — never blocks
Custom resolution and seed control
Direct asset URL on completion

Why agents need image generation

Modern AI agents are increasingly asked to create visual content — product mockups, social media graphics, blog illustrations, concept art, and presentation images. Without an image generation API, agents can only describe what an image should look like.

Flexura's Image Generation API gives any AI agent the ability to actually create images from natural language descriptions. Users simply say 'generate an image of a sunset over Mars' and the agent handles the entire workflow automatically.

How async image generation works

Unlike search, image generation is an async process. When your agent submits a prompt, the API immediately returns a job_id and queued status. The agent then polls the job status endpoint every few seconds until the image is complete.

This async pattern is critical for AI agents because it prevents blocking. While the image is being generated, the agent can continue responding to the user, process other tasks, or provide status updates.

Once complete, the response includes a direct asset_url where the generated image can be downloaded or displayed. Images are stored on Flexura's servers and accessible via a public URL.

NVIDIA FLUX model capabilities

Flexura currently offers the NVIDIA FLUX model for image generation, which excels at photorealistic imagery, stylized illustrations, product photography, portraits, and creative concept art.

You can control the output resolution (default 1024x1024), set a specific seed for reproducible results, and adjust the number of inference steps for quality vs. speed tradeoffs.

Works with every major AI agent

Install the Flexura skill with one command and your agent automatically knows how to generate images. The skill teaches the agent the complete workflow — discover available models, submit a job, poll for completion, and return the result to the user.

Supported agents include OpenAI Codex, Cursor, Windsurf, Claude Desktop, Aider, and Cline. All use the same API key and the same endpoints.

Frequently asked questions

How long does image generation take?

Most images are generated in 3-8 seconds using the NVIDIA FLUX model. The agent polls every 5 seconds and typically gets results within 1-2 polls.

What image formats are supported?

Generated images are returned as PNG files by default. The asset_url points directly to the image file which can be downloaded or embedded.

Can I control the image resolution?

Yes. Pass width and height in the options object. The default is 1024x1024. NVIDIA FLUX supports various aspect ratios.

Is there a limit on image generations?

Free tier users get 5 generations per day. Pro and Business tier users get higher limits. Check the pricing page for details.

Ready to add this capability to your agent?

One command. One API key. Works with Codex, Cursor, Windsurf, Claude, Aider, and Cline.