The persistence layer is probably more valuable here than the visual canvas itself. When running multi-agent workflows on complex tasks, the hard part isn't the agents working in parallel — it's maintaining coherent state when something fails mid-run or needs revision. A canvas that preserves intermediate results and lets you surgically re-run specific branches without starting over would solve a real pain point. Curious how Spine handles partial failures — does the whole workflow restart or can it pick up from the last valid checkpoint?
Spot on. The persistence layer is a huge part of what makes the canvas work.
For failures, we handle it at multiple levels: first, standard retries and fallbacks to alternate models/providers. If that fails, the agents look for alternate approaches to accomplish the same task (e.g. falling back to web search instead of browser use).
For completeness, you can also manually re-run or edit individual blocks if they fail (though the agents may or may not consider this depending on where they are in their flow).
