Hi everyone! I’m noticing inconsistent response times specifically when using Autonomous Nodes. The "typing" indicator (the three dots) shows highly variable latency before appearing. Sometimes the response triggers instantly, but often there’s a delay of 3 to 10 seconds before the typing state even begins. Since this is happening within an Autonomous Node, I’m trying to identify if the bottleneck is the model's reasoning time, the complexity of the task instructions, or a platform-side execution delay. Has anyone else experienced significant lag specifically with autonomous nodes lately? Would like to know if there are any best practices to reduce this "pre-typing" wait time.

Hey! @flaky-printer-85634 which LLM are you using?

Hi @loud-crowd-60573 ! The recommended one. GPT-4.1

Can you try 5.2 and let me know if you see any difference?

Thanks for the suggestion! I already tried switching models as recommended, but unfortunately, the latency persists. I’ve noticed that this isn't exclusive to AI nodes; it happens with every message across the flow. Interestingly, the delay only occurs in the published bot (both on the shared Botpress test link and the client's production site), while the Studio Preview works perfectly and remains fast. Since it happens even with static text nodes, it seems like a platform or integration-side latency rather than a model reasoning issue. Any other ideas on what could be causing this discrepancy between the Preview and the Live version?