Long time no see. I’m with Bubble 6 years and now I’m facing a major blocker with the hardcoded 5-minute workflow timeout, and I know I’m not the only one who suffered from this roadblock. This limit feels increasingly outdated as we move into the era of AI and heavy data processing.
The Problem: I’m building Aicado, where you can create AI Agents and connect them with external databases, MCP’s and each other. This gives us a great power but we have a MAJOR roadblock at that point. With "Deep Search"and “high thinking” capabilities (like GPT5 or Gemini 3 pro) by design, these models take time to “think” to produce high-quality outputs often, this exceeds 300 seconds.
Currently, Bubble treats this valid processing time as a failure, kills the workflow, and leaves my users and me with nothing.
Why we need a change now: It’s not just about AI. Fetching large datasets from external DBs (like Typesense) also frequently hits this wall.
I’ve seen many developers over the years discuss moving to external backends (like Xano or AWS Lambda) or even leaving Bubble entirely just to bypass this single limitation. I really don’t want to fracture my stack or move logic off-platform for something that should be a simple setting.
Please allow us to configure the Workflow Timeout Duration (e.g., in the Advanced settings). Even if it requires a warning or is tied to a specific plan, having the option to extend this to 10-15 minutes is essential for modern “Agentic” workflows.
We need Bubble to scale with the AI capabilities, not cap them.
If you agree with me, please reply this post and upvote the related ideaboard feature from here.
I’ve been frothing at the mouth asking them to fix this for months but they would much rather spend the time enhancing their own AI functionality (the AI agent) rather than empowering their users’ AI functionality.
They don’t actually know what problems a current 2025 app is facing. They’re attempting to build a 2025 app to solve the problem of creating a 2015 app, but no one is trying to build a 2015 app. Those have already been made and are established in their niches. The only way for a new Bubble app to grow (and Bubble to grow by proxy) is to innovate with AI, which Bubble currently handicaps you from doing.
I know it’s not an acceptable solution for something that should be in the core product, but for anyone that does need it, dedicated servers let you configure the timeout.
We ended up needing to spin up our own server to handle this external from bubble. Now this server acts as a middle step between the AI endpoint and the bubble app. This work has actually moved us closer and closer to moving the entire app logic and design out of Bubble entirely.
On a purely strategic basis, you don’t want me to take 2 weeks to set up a FaaS and do a hacky workaround to get past this issue. The#1rule of platform growth is to keep users from leaving your platform. YouTube and other sites have warnings that pop up saying, “Are you sure you want to leave?” X/Twitter deboosts links. Apple is well, you know, Apple. The more you incentivize users to use other services to do basic things, the more you encourage them to move off your platform for good. It’s a slippery slope. It starts out as a Cloudflare Worker, then Xano, and by the time you know it, the person has moved off Bubble for their new builds.
Hey guys @brian.levee@josh@fede.bubble@nick.carroll , you might want to check out this problem. It acts as a major roadblock for those who want to create decent AI functionality in their apps, such as deep search models and MCPs. There is so much frustration around this topic, and sadly, it seems there is no solution other than migrating away from Bubble. At least you can say an estimation or clear answer maybe
btw I noticed that submission in the ideaboard is from 2023
I know it sounds duplicate but it might be a good idea to resubmit as a new idea with all this AI context you shared
Don’t rely on ideaboard too much. I don’t know if it is done by companies but I would like some sort of notepad, a feedback section, I would like to add to that when I face problems. That way you can get continuous feedback about what is wrong with the product. I want 100% guarantee that team will read it.
might be a naive question since I haven’t worked with MCP servers before but can’t you set a WF to queue up the work, and another to receive the results, like a webhook? Do you have to do it all in one?
Receiving an AI response asynchronously via webhook is possible but most services don’t support it and it’s insanely cumbersome when the easy solution is just increasing/configurable timeout.
The number of open sockets on a machine is a really constrained resource, there’s a limit per IP, each one takes memory, and servers need sockets for everything else . It’s very easy with long running processes to blow up a server. Databases and long running queries is a classic.
So I can understand why Bubble doesn’t just swap the timeout from 30 secs to 300 secs. On dedicated - you just blow your self up! Not take down the server.
I guess the work around in Bubble would be to connect from the browser ….
Written in the spirit of the old forum days of trying to find work arounds and tricks, not just moaning about Bubble!
As @georgecollier said its not a exact solution for some cases and its hard to maintain. Also, its not only about MCP’s. For example when I process a file or flow in GPT-5.1 - high thinking it generally passes 300 seconds for just one call and I couldn’t get the result that’s why I can’t use the built in tools (code interpreter, deep search etc.) or functions of an AI model. I can set up a workflow which contains a webhook to get the result at once but it breaks my whole streaming flow & user experience. Right now, the only solution to shift database into external tool (too much hours and hard to maintain - besides if I won’t use Bubble as a fullstack why would I use as a frontend solution?), shift to dedicated (too much money for a small SaaS or a feature) or ask from you for custom timeout value
I’m glad to see you again @lindsay_knowcode this topic gives me a good vibe too, like old forum topics Yes, as you said its hard to open an access to all users but they might include this feature into Growth or Team plan maybe? It seems 300 sec is a bit outdated tho
