How can we help? 👋


Have a look here to find answers to some of the common problems!

The most common advice to any agent issue is: - Forgive & try again - Review the prompt - Review enabled plugins

I get no results for my agent

Sometimes the agent may say that it has successfully completed the task but no results are shown in the “Results” section. Agents can get confused sometimes and we’re actively working on improving this behaviour. If your agent reports task completion without results, we advise to:

  • Review the process step by step to see if the results have been shared in one of the steps
  • Restart the agent with the same configuration

The agent tries to use a plugin which does not exist

You may see the agent try to use plugins which do not exist. Such tools are marked in the process output in square brackets like this: [whistlebells]. Using a plugin like this will fail and likely confuse the agent even further, leading to error state or wrong output. If you see incorrect plugin usage it’s best to cancel the run and try again!

Why does this happen? The most common reasons are:

  • Choosing too many plugins. We advise selecting only one or two specialized plugins.
  • Providing confusing prompt about plugin usage, for example mentioning actions without enabling plugins for them.

The agent tries to use a plugin but it results in an error

We currently have over 300+ plugins available on Spell. These are mostly third-party integrations with external systems which can underperform, get outdated or fail. We’re working on improved curation and review process to make sure each plugin is always up to date and works seamlessly. If you notice an underperforming plugin, please help us by reporting it to 🙏


My agent has stopped due to maximum iterations. What does it mean?

Each round of thinking and doing by an agent is called an iteration. When creating an agent, you can configure the number of iterations to limit “how far” you want to allow the agent to go in its task. This is done to prevent excessive quota usage; however, if you set the number of max iterations too low for your task, the agent might be not able to complete it - and then it will stop.

If your agent has stopped due to reaching the maximum iterations, you can restart it with the same configuration but increase the maximum iterations.

At the moment the maximum number of iterations that can be set for an agent is 15. Please contact our support team if you wish to increase this number.

I have subscribed but can’t access the app

Sorry to hear that you are blocked from accessing the app! Let’s get this resolved ASAP: please reach out to and a human team member will fix this for you.