Image generation

Generating an image

Send a prompt to the images endpoint. The response includes a URL (or base64 payload) for each generated image.

from tnsa import TNSA

client = TNSA()

result = client.images.generate(
    model="IGen-1.5",
    prompt="A watercolor fox sitting in a misty forest at dawn",
    size="1024x1024",
    n=1,
)

print(result.data[0].url)

Parameters

• prompt — a text description of the desired image.
• model — an IGen model id, e.g. IGen-1.5.
• size — output resolution such as 512x512 or 1024x1024.
• n — number of images to generate (1–4).
• response_format — url or b64_json.

Prompting tips

• Describe subject, style, lighting and composition explicitly.
• Use IGen-1 Nano or IGen-1 Fast for rapid drafts, then re-run with IGen-1.5 for the final.
• Iterate: small wording changes can meaningfully shift the result.

Browse all available image models on the Models page.