Every plan includes the agent. Free and Cadet plans have usage limits that refill over time. Voyager and Enterprise plans include monthly credits, more advanced models, and top-up credits. See Plans and pricing.
Opening the agent
Click the Agent tab in the left sidebar, or pressCmd + 5 (macOS) or Ctrl + 5 (Windows).

- Chat dropdown: Switch between your chats.
- +: Start a new chat. Your other chats are kept.
- Panel Options (•••): Close the current chat, close the other chats, clear all chats, or open the chat in the stage area.
Opening the agent in the stage area
For more room, open a chat as a tab in the stage area, next to your artboards:1
In the Agent tab, click Panel Options (•••).
2
Choose Open in Viewport.
Writing a prompt
Type what you want in the prompt box at the bottom of the panel and pressEnter. Be specific about what should change and how it should behave.
Create a script that draws a rounded rectangle. Width, height, color, and corner radius should be inputs. When the inputs change, the drawing should update.The agent replies with a summary of what it changed, shows each step it took, and explains anything it created, such as a new script in the Assets panel. You can also attach an image to your prompt, such as a reference or a sketch. Not every model supports images.
Ask and Build
Choose a mode below the prompt box:- Build designs, animates, and writes code in your file. It uses credits faster.
- Ask answers questions about your file or about how to use Rive, without changing anything. It uses fewer credits.
Model and effort
Choose the model and the effort level, Small, Medium, or Large, below the prompt box. Higher effort can handle bigger requests, but uses more credits. Advanced models are available on Voyager and Enterprise plans. On Voyager and Enterprise plans, admins can add your workspace’s own API keys for select model providers, so usage is billed to your provider account instead of your Rive credits. Choose API keys from the model menu.What the agent can do
The agent works in the open file. It can:- Design: Create shapes, paths, and vector artwork, and import SVGs.
- Build layouts: Create layouts and groups.
- Use components: Create and arrange components.
- Rig: Add bones and meshes.
- Animate: Create timelines and keys, state machines, and listeners.
- Bind data: Create view models, properties, and property groups.
- Write code: Create and edit scripts and shaders, run tests, and read the console to fix errors.
- Manage assets and tags: Work with the file’s assets and tags.
- Inspect the file: Find objects and read their properties, so it can answer questions about the file.
Credits
The agent is billed in AI credits, which are expressed in USD: $1 in credits is $1 of AI usage. Each paid seat adds monthly credits to your workspace:- Voyager: $20/seat/month
- Enterprise: $40/seat/month
- Monthly credits reset each month and don’t roll over.
- Monthly credits are always used before top-up credits.
- Top-up credits don’t expire.
Adding credits
Workspace admins on Voyager and Enterprise plans can buy top-up credits.1
On the Home tab, click the dropdown next to the workspace name and choose Admin.You can also click Usage on the Agent credits card.
2
On the Overview tab, find the Top-up credits card and click Purchase credits.
3
Choose an amount and complete the purchase.
FAQ
Is my code or project data used to train the AI or any other LLMs?
Is my code or project data used to train the AI or any other LLMs?
No. We have a Zero Data Retention policy active with our LLM provider. Neither we nor the provider use user data to train AI models.
Can I disable the agent for my workspace?
Can I disable the agent for my workspace?
Admins on Enterprise workspaces can turn the agent on or off for the entire workspace at any time.
Will the AI make mistakes?
Will the AI make mistakes?
Yes. Like any AI system, the agent can make incorrect or suboptimal changes, though it tries to check and correct its work. Always review what it creates, especially code.
Can I provide feedback on the AI's responses?
Can I provide feedback on the AI's responses?
Yes. Please send feedback to our Community.