Canva launched two significant backend AI features aimed at enhancing its integration with AI chatbots. Although these tools aren’t visible to users on the frontend, they mark a major step toward smarter, more connected workflows across platforms.
The first is Canva’s Deep Research connector for ChatGPT, which allows OpenAI’s chatbot to access and understand user content directly from Canva projects. With this integration, ChatGPT can answer questions like “Summarize our Q1 campaign strategy,” by locating the relevant deck in Canva and generating a response based on its content. This eliminates the need to manually upload files or switch between platforms for assistance. The feature works when ChatGPT’s Deep Research mode is active and is available to users across all subscription tiers.
The second major addition is an open Model Context Protocol (MCP) server, enabling third-party AI assistants to perform tasks within Canva. Developed as a standardized connection protocol by Anthropic, MCP allows chatbots like Google’s Gemini or Anthropic’s Claude to execute commands such as generating pitch decks, resizing files, or importing/exporting assets—all directly within chat interfaces.
For example, a user could instruct Gemini to “Create a presentation based on this chat,” and receive a complete Canva deck instantly, without leaving the conversation.
While the MCP server is now live, Canva says integrations with platforms like Salesforce, ChatGPT, Claude, and others are in the pipeline, though specific timelines haven’t been disclosed.