Summon's Image tab creates new images from a prompt and edits existing ones, including the frame under your playhead, without leaving After Effects. Results come back into your project, ready for the timeline or as references for video.
In the full tutorial, the playhead is moved to a shot, From Playhead captures it, and a character sheet is added as a second reference to ask for a front-facing angle. Because the capture is what the composition shows, the next step uses the framing you've actually built in After Effects, not the untouched source image.
Three buttons sit at the right of every prompt box: EN translates the prompt to English, which most models follow far better; ✦ translates and enhances it for the current model; ↩ restores your original. Translate and enhance are small language-model calls billed by fal.ai, and they cost very little.
The Image tab has 15 built-in models, including Nano Banana Pro and Nano Banana 2, GPT Image 2.5 (Sunburst and Flare, each with an Edit version), Seedream 5.0 Pro and Lite, FLUX.2 [pro], Ideogram V4, Recraft V4 for design, Grok Imagine, Kling Image O3 Edit, and Flux Schnell for fast, cheap drafts. More can be added with the Model Manager.
An image is the cheapest way to try a model and check a prompt before you spend on video. In the full tutorial, a 2K character sheet from Nano Banana Pro had a recorded cost of 15 cents.
Upscale an image locally through your installed Topaz apps, free with your existing Topaz license, or in the cloud on fal.ai. Upscales appear in History, where filters tell the original from the enlarged version.
Like every generation, images run on your own fal.ai key at fal.ai's listed prices, with no Summon markup; the price per image is shown next to each model's name. How the pricing works.