AI generation overview

AI generation is how you produce media inside AdFactory — turning a prompt or a reference into images, video, voiceovers, music, lipsync, and captions. It all happens in the creative editor, and every result becomes a material you can refine and ship.

The creative editor is a step-by-step flow, with each kind of media as a tab across the top

Where generation happens

You generate inside a creative, which you open from a task. The editor is a step-by-step flow, and each step is a tab across the top.

If you're producing the creative, your media steps are, in order:

1. Picture — generate images.
2. Audio — generate voiceovers.
3. Video — turn images and frames into motion, and upscale clips.
4. Music — generate background tracks.
5. Submit — send your work for review.

Reviewers see additional steps after these — Review Materials, Select Body Clips, Subtitles, Upload Final Media, and Final Review — for finishing and approving the creative.

On each media step you'll see a form on the left and a live panel on the right showing everything you've generated of that type so far. New results appear in that panel as they finish.

The typical chain to a finished video

The steps are designed to build on each other. A common path from nothing to a polished video looks like this:

1. Start in Picture. Generate one or more images as your base.
2. Bring it to life in Video. Use Image to Video to animate a single image, or generate two keyframes and stitch them together with Frames to Video. You can also generate motion straight from a prompt with Text to Video, and sharpen any clip with Upscale. See generating and upscaling video.
3. Add sound. Generate a voiceover in Audio, and create a track in Music (including instrumental-only options).
4. Sync the face. Use Lipsync to match a talking-head clip to your voiceover so the mouth movements line up.
5. Caption it. Add subtitles so the message lands even with the sound off.
6. Submit the creative for review.

You won't always need every step — a static image ad might stop after Picture — but this ordering is the backbone of most video creatives.

How each generation works

On every media step, pick a model, write a prompt, choose how many outputs, then start the job

The pattern is the same everywhere:

1. Pick a model for that generation. The available modes and options depend on what the model supports — anything it can't do is simply hidden or grayed out, so you only ever see valid choices. Use the ⓘ icon next to the model to open its details — a short description of what it's best at, plus model-specific prompt tips to help you get a better result.
2. Write your prompt and set the controls — things like aspect ratio, reference media, or quality. A clear prompt is the biggest driver of a good result, and prompt optimization can tighten it for you.
3. Choose how many outputs you want — up to four variations from a single prompt in one batch.
4. Start the job. Every generation runs in the background, so you can keep working while it processes.

Jobs, results, and cost

Each generation runs as a background job. Results show up in the live panel as a grid of thumbnails — from there you can keep the ones you like and discard the rest. To explore a different look, request more variations without rewriting your prompt.

Finished results land in the right-hand panel — this batch card holds all four variations from one prompt, with the prompt and model shown beneath

Before you start, AdFactory shows an estimated cost in credits; once the job finishes, you see the actual cost. Roughly 100 credits equals about $1. For how usage is metered and how to top up, see credits, top-ups, and payments.

Every finished batch card shows what it cost — the credits spent and how long the job took

Tips

• Generate a small batch first to dial in your prompt, then scale up once you like the direction.
• Combine generated media with uploaded files when you already have assets to work from.
• When the creative is ready, Submit moves it into review for approval.