From 285 credits/genView API documentation
Explore the AI Effects API through visual examples. Choose a template, compare sample inputs and output, and review request code before integrating the effect.
Quickstart
Search using a full or partial template identifier. Search ignores capitalization, whitespace, and hyphens. Compare the saved result with its sample inputs. For multiple images, check the role of each one: switching their order can tell a different story even if both files are valid portraits. Consider where users see the result. A video might look good in a player but lose its subject when cropped into a narrow feed. A still image might work better if users need something they can inspect at a glance. An AI Effects API sample illustrates one result with particular inputs; it does not guarantee an identical result for every upload.
Open the documentation linked to the selected template. Locate its identifier, required fields, accepted media, and optional settings. Compare those fields with the saved example before replacing the sample assets with your own. Keep the original field names and preserve the order of any inputs whose roles differ. If an example is available in your preferred request language, copy it as a starting point. When no matching example is available, construct the request from the template documentation. Borrowing an entire payload from a similar effect can leave out a required asset or introduce an unsupported setting. The AI Effects API contract for the selected template is the reference for both the request and the form you build around it.
Follow the authentication instructions and submit the request using the selected template. Keep the returned task identifier with the corresponding inputs so that progress and results stay attached to the right item in your interface. Submission starts generation; a successful submission alone does not mean the media is ready to display. While the task is processing, preserve the chosen effect and show its current state. When it completes, read the documented output and present it in the appropriate image viewer or video player. Handle failure separately so an unfinished or failed task never looks like an empty success. Before promising a permanent download, read the asset guidance for the AI Effects API and decide how your application will retain the completed file. Give a completed video a player with pause controls, and present a still image at a size that makes the transformation easy to inspect. Keeping the original input beside the finished result can help your users decide whether to keep the output or try different assets. This also helps distinguish a poor source photo from a framing problem in the final display area.
Template guide
Compare published video and image templates for the AI Effects API.
Inspect subject motion and transition timing in the sample clip. Watch the subject through the entire clip. Compare the opening and closing frames with the source images, especially when a transition moves between two people. This shows which portrait belongs in each input. An AI Effects API cover illustrates a single moment; play the clip at its intended size to judge its timing.
Check portrait framing and composition. Compare the source portrait with the finished still. Look at which features remain recognizable, where the subject sits in the frame, and how the background changes. If the result will appear in a narrow feed, check that the crop keeps the face and the effect visible together.
The gallery displays saved examples. Opening a card in the gallery lets you inspect the output, its sample inputs, and the associated documentation. It does not submit a generation request. This makes it a useful place to compare candidates before you build an upload flow or write request code.
To generate a result with your own media, choose the template and follow its AI Effects API documentation. Keep example content visually distinct from user results in your application. A preview shows what the effect did with the published assets; the generated output will depend on the assets your user supplies.
Build the upload form from the specific template requirements. Count the required inputs, identify their media types, and give each field a label that explains its role. If the example uses two people, explain which person belongs in each position. Keeping the example close to the fields is more useful than asking users to infer the sequence from file names.
Optional controls also vary. Before showing a prompt box, duration slider, or resolution menu, check which values the template accepts. An AI Effects API request should contain the fields that belong to its selected effect. Changing templates may require changing the form even if both templates produce the same media type.
For a video template, check the documented duration and resolution before adding a settings panel. Offer a duration selector only when the template accepts that field. For image templates, check the number of subjects as well as the number of files. An effect that combines two portraits may need a separate upload field for each person.
Use the AI Effects API quote endpoint for the selected template with the parameters you plan to submit. The estimate belongs to that combination of effect and settings. A price shown for another template is not a reliable estimate, even when both produce a video or accept the same number of images.
If users can change a setting that affects price, obtain an estimate for the revised request before they generate. Place the estimate near the generation action so they can relate the credit amount to the inputs they have selected. Consult the AI Effects API documentation for the template you are quoting rather than deriving a price from the appearance of its preview.
Choose an AI Effects API template when its example demonstrates your desired result and its inputs fit the assets your product collects. For example, a portrait upload flow is easier to integrate when the chosen effect expects exactly the portraits your users already provide. Check the contract before committing to the layout.
If you need a different kind of input or a control the template does not expose, compare the image or video model documentation. The AI Effects API is a useful starting point for a repeatable visual treatment. A general model may be a better fit when your users need to define the treatment themselves.
The gallery pairs published template documentation with visual examples. Each example gives you a saved output to compare with its inputs. Newly published documentation appears first, and the available mix of video and image effects can change as templates are published.
When revisiting an integration, look up the template by its identifier instead of relying on where a card appeared in the gallery. Return to its AI Effects API documentation to check the request and supported settings. A familiar cover image can help you recognize the effect, but the identifier is what connects your application to its documented template.