SuperAPI AI Directory

Figma Design agent

The October 6 release adds file search in chat, library guidelines and live visibility into teammates’ agents. Library authors can supply Markdown guidelines, and enabled libraries provide that context when prompting. File attachments and selected nodes help constrain a request. Figma’s GA release n

Open website

About this tool

Figma Design agent is an AI agent launched by Figma and officially released after ending its beta on October 6, 2026. It is primarily aimed at design workflows, helping users perform operations such as searching, referencing, generating, and editing in design canvases and files.

[Purpose and positioning] This agent is embedded in Figma and can be used as an AI assistant in design collaboration and creative workflows. According to the official release notes, product designers can use it for ideation, speeding up work, and creative expression. Users can give the agent tasks, such as searching for files within the platform, inserting or referencing design assets, or performing image editing on the canvas.

[Main capabilities] The main improvements and new capabilities of the official version compared with the beta version include:

  1. Reduced latency and improved response speed.
  2. Search Figma files in the conversation: you can prompt the agent to find files across the entire platform.
  3. Support for inserting or referencing various types of Figma platform content, including Figma Design design files, FigJam sticky notes, Figma Slides content, and more.
  4. Design library guidelines: you can add guidelines to a Figma design library, that is, upload Markdown files containing rules, best practices, anti-patterns, and other content; when the user enables that library in a prompt, the agent reads these files every time to gain context about the design library.
  5. View teammates' agents in real time: you can watch teammates' agents work on the canvas in real time, and click the chat stack to view or follow along.
  6. Generative plugins are officially available: you can choose an image model and use the agent's entry point on the canvas for image editing.
  7. Custom skills: you can use the agent to brainstorm, create, and publish custom skills; custom skills are already available in the starter plan.

[How to use]

  • Click the "+" button in the lower-left corner of the prompt box, then click attach Figma files to add context from the Figma platform.
  • You can copy and paste Figma nodes directly into the prompt box to provide context and reference images.
  • To use design library guidelines, first add Markdown guideline files to a Figma design library, then enable that library when prompting.
  • To use the Gemini 4 Argon model (release notes dated October 13, 2026), open the model selector in the Figma Make chat box and select Gemini 4 Argon; this model is available in all plans.

[Applicable scenarios]

  • Ideation, creation, and iteration during the product design process.
  • When there is a need to reference existing design assets, sticky notes, or presentation content as context.
  • Viewing teammates' agents working in real time during team collaboration.
  • Constraining generated content based on design library specifications (guidelines) to align with the team's design system.
  • Performing image editing on the canvas or using generative plugins.

[Limitations and billing notes]

  • Starting October 6, 2026, Figma agent uses AI credits.
  • Specific availability by plan, credit quotas, pricing, and other information are not provided in the materials and need to be checked in the official help center.
  • Other features such as Motion being in public beta and multiple GitHub organization connections being limited to Organization and Enterprise plans are other product updates from the same release notes and are not necessarily directly related to the Figma agent itself.