Franchement, c’est n’importe quoi. Je le teste depuis 6h du matin et je n’ai toujours pas réussi à construire un seul workflow. J’ai pourtant tout fait correctement : clé API Google, clé API OpenAI… même résultat à chaque fois. Ça charge, puis plus rien. Beaucoup trop de buzz pour très peu de résultats. J’ai même demandé la structure de prompt recommandée pour l’aider à accomplir la tâche correctement, mais ça ne donne absolument rien. Une vraie perte de temps, à éviter !
---
Honestly, it’s a joke. I’ve been testing it since 6 a.m., and I still haven’t managed to build a single workflow. I did everything right—Google API key, OpenAI API key—same result every time. It loads, then nothing. Way too much hype for such poor results. I even asked for the recommended prompt structure to help it complete the task properly, but still, absolutely nothing. Total waste of time—don’t bother!
n8nChat
Launch day is finally here, just 30 days after I got started on this project 😁
Started as a tool I wish existed for n8n, so I just decided to build it for myself. Turns out a bunch of other people want it too... over 2,000 on the waitlist!
Conduit
Does n8n chat help build out better chat experiences?
Would love to chat with anyone who's built production-scale conversational workflows and learn from you!
I'm thrilled by how AI streamlines workflow setups in n8n! 👀
Love it! I'm using n8n a lot, this would be a game changer!
n8nChat
New guide for a common n8n pain: Loop Over Items / Split in Batches still OOMing (including silent Docker heap deaths).
Pattern that works: Loop Over Items → Execute Sub-workflow → tiny Set return from the child. Batching alone does not free parent RAM.
https://n8nchat.com/blog/n8n-loo...
n8nChat
Second tutorial is live: n8n webhook response modes (Immediately vs When Last Node Finishes vs Respond to Webhook), and why callers hit a ~100s timeout when the Agent or a slow HTTP Request sits on the critical path.
https://n8nchat.com/blog/n8n-respond-to-webhook-timeout
n8nChat still does one job: open your workflow as context and place/fix the graph on the official n8n canvas. This is the how-to, not a new product feature.
n8nChat
Shipped a short guide on the n8n Wait node and webhook resume URLs — the gotchas that bite when you use time delays, forms, or unique resume URLs. Live here: https://n8nchat.com/blog/n8n-wai...