# Creation templates

> Use image and video templates in a new chat, bring templates in from the website, create personal templates from reference images, and edit video template workflows.

- Source: https://beav.pro/en/docs/workflows/templates

Checked 2026-10-07. Applies to Beav desktop 2.8.13.

Templates give the AI a fixed visual direction and input requirements, such as "Big-text portrait · 3D composition cover" or "Home · White-line hand-drawn diary". Public templates come from Beav's online template marketplace and are not bundled with the installer. The first use needs an internet connection to load them; after that, Beav shows the local cache first.

## Use a template in a new chat

1. Open a new chat. The template list is below the input box. Filter by industry (**Personal brand**, **Tech creator**, **Food & dining**, **Real estate**, **Home**, **Pets**, **Beauty & care**, **Outdoor**) and by type (**Cover**, **Image**, **Xiaohongshu**, **Video**) at the top. Templates must match both filters.
2. Hover over a card and click the full-screen button to see the template details: preview, categories, creation method, and reference image requirements.
3. Click **Use template**. A small card labeled **Template** appears above the input box.
4. Add your topic and requirements, upload reference images as the template describes, then send.

![New chat page with industry and type filters, a Create template card and template covers below the composer](https://beav.pro/screenshots/new-chat-templates-en-2026-10-07.webp)

*Creation templates sit below the new-chat composer, filtered by industry and type.*

Choosing a template does not submit a task; creation starts only after you send. A template marked as needing reference images cannot be submitted without them. The template's sample images are only previews of the result and are not sent as your references. Using another template replaces the current choice. Removing the template card does not affect the text and attachments you have already entered.

After you send, a small card with a cover image showing the template you used appears below the message.

## Bring a template in from the website

On the website's [templates page](https://beav.pro/en/templates), you can browse and search all public templates and see their full prompts and reference image requirements. Click **Use in Beav** on a details page and allow the browser to open Beav. The template is placed in the input box of a new chat; add your content and send it yourself. Nothing is generated or charged automatically.

## Create a personal template from a reference image

In the new chat's template list, click the first card, **Create template**, which shows a plus sign. Drag in or choose an image, preview it, and click **Create**. Beav breaks the image down in an ordinary chat and saves it as a personal image template in the current workspace. The AI proposes the name, and you can keep revising it in the same chat. Personal templates belong only to the current workspace and are not published to the public marketplace.

Creating a template and generating images with it are separate steps: after the template is saved, you still need to select it, fill in the inputs, and send before any image is generated.

## Video templates

Choose a video template in the template list, add your topic, requirements, and any assets it needs (for example, product images), then send. Beav opens the template's generation dialog and starts generating according to its workflow. The dialog shows each node's status (**Waiting for upstream**, **In progress**, **Done**, **Failed**, **Needs attention**) and intermediate outputs:

* Click **Retry node** on a failed or cancelled node.
* Click **Cancel generation** while it is generating.
* When a video project is generated, click **Edit video** to keep working on it in the editor.
* **Edit workflow** at the top of the dialog opens a separate workflow editing window.

You can also skip templates and describe the video you want directly in a new chat; the AI generates it with the current video model.

### Edit a video workflow

The workflow editing window contains nodes and connections:

* Double-click or right-click an empty area of the canvas to add a node. With a node selected, click the **+** next to its output to connect the next step directly.
* The bottom toolbar switches between the **Select tool (V)** and the **Hand tool (H)**; **+** opens the node catalog, where you can search for nodes. After a box selection, you can copy or delete. `Cmd/Ctrl+D` duplicates the selected nodes.
* The right side shows the **AI assistant** by default. Describe the steps, parameters, or connections to add, for example "Add a prompt node and connect it to Generate video and Final output." Use the **+** in the input box to choose account profiles, knowledge base videos, or local files as references (up to 12 at a time), and the AI extracts their common patterns and turns them into a template. If you only want to discuss a plan, say "don't apply it yet."
* Switch to **Inspector** to configure the selected node, such as the generation model, voice, speaking speed, size, and duration.

AI changes to a workflow do not start generation automatically. To do a trial run, still use the run button.

## What generation requires

Image and video generation uses credits or external services you have configured. Available parameters (duration, aspect ratio, number of reference images, whether there is sound) vary by model; follow what the current interface shows. A successful submission does not mean generation is finished. Do not resubmit just because there is no result yet; check the task status first.