Figma adds voice prompts, version history and question cards in Make
Figma added voice-to-text prompts, version history, question cards and clearer context in Make, while Weave gained node unpacking, iterators and flow-preserving delete. The updates make it easier to capture prompts and manage asset-heavy AI flows inside Figma; try the new controls if your team needs traceable edits.
TL;DR
- Figma's Make update post and Figma's release notes shipped four new chat controls at once: voice-to-text prompts, version history, question cards, and Clear Context.
- According to Figma's demo thread, Make now lets users dictate prompts, while the release notes say the text is cleaned up before submission.
- figmaweave's Weave demo added three workflow edits for asset-heavy graphs: Unpack Nodes, iterators, and Shift + Delete that preserves the rest of the flow.
- Figma's FigJam MCP post also expanded the company's agent tooling with generate_diagram, figma-use-figjam, and get_figjam, which are described in Figma's FigJam MCP post as ways to turn docs and codebases into editable boards.
You can skim Figma's release notes, open the new Clear chat context help doc, and browse Figma's FigJam MCP writeup. The useful pattern across all three updates is less prompt re-entry, less graph cleanup, and more ways to keep AI work inside the same canvas.
Make chat controls
Figma bundled four controls into Make's chat loop instead of shipping them as separate experiments. In the release notes, the company describes them like this:
- Voice-to-text: dictate prompts in chat, then review cleaned text before sending.
- Question cards: pick from structured options with short tradeoff descriptions before Make continues.
- Version history: track every build state and jump back to an earlier one instantly.
- Clear Context: reset the session from settings in one click.
The new Clear chat context help doc adds the most concrete detail: Make keeps conversation history, attachments, guidelines, and connected tools in memory, and Figma says long or complex sessions can push outputs off track. That makes Clear Context less like a cosmetic cleanup button and more like a manual reset for a drifting build.
Weave graph controls
figmaweave's feature drop framed the update as three small edits, but together they remove the fiddliest part of working with multi-asset nodes.
- Unpack Nodes splits a packed node into individually editable assets.
- Create Iterators turns a multi-asset node into a one-by-one processing loop.
- Shift + Delete removes a node without breaking the rest of the chain.
That lands right after Figma's reference-input post and its Draw update thread, which added more ways to feed source material into image generation: copy and paste, selection-based references, and drag-and-drop files. Figma's AI tools are getting more traceable at exactly the points where creative workflows usually get messy, context in, branch logic, and rollback.
FigJam whiteboard tools
Figma's other April 28 ship was aimed at coding agents, but the mechanic is useful for designers too. In Figma's FigJam MCP post, the company listed three additions:
- generate_diagram for architecture diagrams and ERDs
- figma-use-figjam for adding notes, code blocks, and annotations directly to a board
- get_figjam for reading the board and mapping next steps
Figma's FigJam MCP post says those tools launch on top of the earlier use_figma canvas tool, and adds one extra detail the tweet does not: the team also built workflow skills like generate-project-plan to turn docs, codebases, and conversations into visual boards. That is a different kind of AI surface than Make or Weave, but it points at the same bet, prompt systems are easier to manage when they stay visible on a canvas instead of disappearing into chat history.