Petter Magnusson

Extended thinking is impressive but assumes continuous execution. For workflows needing human approval gates between steps—pause after research, review, resume—you need structured human-in-the-loop workflows with state preservation, not autonomous agents.

Both assume continuous execution. If you need human approval gates between steps, or workflows that pause and resume days later with full state preserved, you need purpose-built tooling for structured human-in-the-loop workflows.

Both assume continuous execution. If you need human approval gates between steps, or workflows that pause and resume with full state preserved, you need purpose-built tooling for structured human-in-the-loop workflows, not RAG frameworks.

Congrats on the launch, @dnzozcan ! One architectural question: for workflows that need human approval gates during execution (not just at planning), how does Maia handle state persistence? Like if someone needs to pause mid-workflow, get colleague feedback on a draft, then resume from that exact point. We've found that's where most autonomous tools break down, the handoff between "AI executes"...

This looks solid! The evaluation-first approach is exactly what's needed. Curious how you handle workflows that need human approval gates between steps, not just eval metrics, but hard stops for review before continuing?

Removing setup friction is underrated work. When people can go from idea to something running without fighting their environment, they spend energy on the actual problem instead of the tooling. That’s usually what decides whether a tool gets adopted or quietly abandoned. Congrats on the Desktop launch.

What stands out to me is the emphasis on making workflows understandable, not just runnable. When people can clearly see how decisions flow and how state evolves, AI becomes something teams can actually rely on day to day. That’s usually where these tools either stick or fade. Nice work. Curious to see how this holds up as usage grows.

Congrats on the launch. Tools that make execution visible tend to earn trust faster, especially once workflows get more complex. Interesting to see that baked in here from the start.

Interesting. In practice we’ve found structure matters more than “smarter prompts”. Workflows with explicit steps + human checkpoints tend to ship better than fully autonomous flows.

Very cool to be able to build so fast! How does this position relative Lovable etc?

Would love to get some smart dashboards going for purposewrite.com !

OMG almost #1 now!!! Thanks for all the support! Me and my team will thank you all by making even more fantastic Guides (mini apps) inside purposewrite. Lets fight the generic pointless ai content now!

Looks very cool, but I did get stuck in the UI a couple of times, maby sit next to a noew user and see what happens?

Frustrated by Generic AI Content? So was I. Hey everyone! I'm thrilled to finally share purposewrite with you all. As someone who runs a content production company, I was getting increasingly frustrated by the flood of generic AI content out there. You know, the kind that adds no real value to the reader or the poster. I wanted something different—something that helps you come up with genuinely...