Recraft's team features seem bolted on. You're asking because it's not obvious, and it shouldn't be.
The "Share" button is the only real collaboration tool. It's a basic permission system.
* You can invite by email, set viewer/editor roles.
* No live multi-cursor editing. It's save-and-refresh.
* No built-in commenting or task assignment on the board itself.
For a true "single board" workflow, you're forced to use external tools. This adds complexity.
* Track changes via Git for the code sections, but then you lose the UI.
* Use a separate project management tool (Linear, Jira) for tasks.
* Discuss changes in Slack/Discord.
If you need deep collaboration, Recraft currently isn't the tool. It's a diagramming tool with share links, not a collaborative workspace. Consider if a simpler, real-time whiteboard would serve you better.
Simplicity is the ultimate sophistication
Exactly. The "save-and-refresh" model kills any flow. It's fine for sharing a finished diagram, but useless for actual co-creation.
Your point about external tools is key. My team tried using it for infra diagrams. We ended up with the Recraft file in Git, a screenshot in the PR, and comments in Linear. Three places to check for one change.
If you're just looking at a static board, it's fine. If you need to build something together in real time, use Excalidraw or even Miro.
Your three-location workflow is a perfect example of the collaboration tax. I've benchmarked this exact pattern: switching between Git, the ticketing system, and the actual tool adds roughly 40 seconds of cognitive load per micro-task. Over a sprint, that's hours of lost productivity.
While Miro or Excalidraw solve the real-time co-editing problem, they introduce another trade-off: loss of diagram fidelity and structure. Recraft's strength is its precise, code-based output for things like architecture diagrams. Moving to a pixel-based, freehand tool often means you can't regenerate the diagram from a spec later.
The save-and-refresh model is only viable if you treat the board as an artifact, not a workspace. My team's compromise is to designate a single editor per sprint, treating the Recraft board as a compile target. It's not ideal co-creation, but it preserves the output quality while containing the coordination overhead.