Agents Experiencing Slow Performance and Unexpected Crashes

Brendan L. · 2025-09-09 20:03:24.143018+00

Anyone else having issues with their Agents performing REALLY SLOWLY. Like getting stuck on simple tasks, or quitting unexpectedly

Agents Experiencing Slow Performance and Unexpected Crashes | Lindy Community

8 comments

· Sorted by Oldest

Ronnie P.
·
·
lots of issues as of late.
Nitai B.
·
·
Not me, I've had my agent save over 88 hours last week and I've spent over 17,000 credits. It's been a life saver.
Ronnie P.
·
·
Ha. That's great to hear Nitai B.. Curious, how many agents do you have in the wild? Are you building for yourself or others?
❤️1
Nitai B.
·
·
I work for a small Medicare insurance brokerage and we have 5 employees. I create all the automations but also do everything everyone else does. This is 1 agent. The next agent is already in queue but I can't do anything with it until after October 15th, when we're able to send out Medicare Plan reviews for open enrollment. The next agent is probably going to use 400+ credits per task, so about $4.00 but it will be worth it, we'll be able to perform hundreds to thousands of hours of work in a few days. I'm only building for my employer right now but I intend to take my automation knowledge to an independent business role.
👍1
Ronnie P.
·
·
Nice one. Sounds like you're up to some interesting stuff.
Nitai B.
·
·
It's just simple computer use for a portal that doesn't' have an API. It's the only way to interact with our quoting tool. Computer use is best used for doing things you can't do with an API or an integration with Lindy.
Ronnie P.
·
·
Yep. I'm using computer use for some QA workflows. Most of my Lindy agents are heavy in APIs though. Whatever gets the job done ✅
Brendan L.
·
·
Glad to hear you're running smoothly Nitai! It's strange - sometimes there are zero issues, and sometimes the agents get hung up on the simplest parts of a task and then quit unexpectedly. I currently have 4 agents that do different tasks, all in a sort of "pipeline". One is a very simple data scraper, another is a researcher that uses the computer, another is an outreach agent. All fairly modest in their capabilities. But I've spent a LOT of credits just trying to get them to run through a task without quitting unexpectedly or getting hung up. Restarting their task from where they left off, etc.
👀1

Ronnie P.
·
·
lots of issues as of late.
Nitai B.
·
·
Not me, I've had my agent save over 88 hours last week and I've spent over 17,000 credits. It's been a life saver.
Ronnie P.
·
·
Ha. That's great to hear Nitai B.. Curious, how many agents do you have in the wild? Are you building for yourself or others?
❤️1
Nitai B.
·
·
I work for a small Medicare insurance brokerage and we have 5 employees. I create all the automations but also do everything everyone else does. This is 1 agent. The next agent is already in queue but I can't do anything with it until after October 15th, when we're able to send out Medicare Plan reviews for open enrollment. The next agent is probably going to use 400+ credits per task, so about $4.00 but it will be worth it, we'll be able to perform hundreds to thousands of hours of work in a few days. I'm only building for my employer right now but I intend to take my automation knowledge to an independent business role.
👍1
Ronnie P.
·
·
Nice one. Sounds like you're up to some interesting stuff.
Nitai B.
·
·
It's just simple computer use for a portal that doesn't' have an API. It's the only way to interact with our quoting tool. Computer use is best used for doing things you can't do with an API or an integration with Lindy.
Ronnie P.
·
·
Yep. I'm using computer use for some QA workflows. Most of my Lindy agents are heavy in APIs though. Whatever gets the job done ✅
Brendan L.
·
·
Glad to hear you're running smoothly Nitai! It's strange - sometimes there are zero issues, and sometimes the agents get hung up on the simplest parts of a task and then quit unexpectedly. I currently have 4 agents that do different tasks, all in a sort of "pipeline". One is a very simple data scraper, another is a researcher that uses the computer, another is an outreach agent. All fairly modest in their capabilities. But I've spent a LOT of credits just trying to get them to run through a task without quitting unexpectedly or getting hung up. Restarting their task from where they left off, etc.
👀1