Skip to content

Support & Community > Troubleshooting and support

Sending feedback and logs

Open in ChatGPT ↗
Ask ChatGPT about this page
Open in Claude ↗
Ask Claude about this page
Copied!

Send Derp feedback, bug reports, and feature requests, and gather logs, crash reports, CPU samples, and AI conversation IDs to attach to them.

  • For subscriber technical issues or questions (bugs, credits, etc.), email support@derp.dev.
  • For subscriber billing issues or questions (refunds, cancellation, etc.), email billing@derp.dev.
  • For enterprise, please direct all feedback and issues to your designated Slack channel.

The /feedback slash command is the fastest way to report a Derp bug, flag a regression, or file a feature request from inside the terminal. It opens the feedback form without starting an Agent conversation or consuming credits.

Whether you use the /feedback slash command or file an issue manually, a good feedback report answers these questions up front:

  • What happened? Describe the observed behavior in one or two sentences.
  • What did you expect? Describe the behavior you expected instead.
  • How do we reproduce it? List numbered steps when possible. If you can't reproduce the issue reliably, mention that too.
  • What version of Derp are you on? /feedback fills this in automatically; for manual reports, copy it from Settings > Account.
  • Logs, screenshots, or conversation IDs. See Gathering Derp logs, Collecting crash reports on macOS, or Gathering AI conversation ID below.

See the Slash Commands reference for the full list of commands available in Derp.

Retrieve Derp's logs by following the instructions for your platform below. Locate the log file and attach it to your GitHub issue or email.

The Derp log files are located at ~/Library/Logs/.

Derp logs on macOS

Run the following to zip the Derp logs to your Desktop:

Terminal window
zip -j ~/Desktop/derp-logs.zip ~/Library/Logs/derp.log*

Derp Preview logs on macOS

Run the following to zip the Derp Preview logs to your Desktop:

Terminal window
zip -j ~/Desktop/derp_preview-logs.zip ~/Library/Logs/derp_preview.log*

If Derp crashes, macOS may generate .ips crash report files in ~/Library/Logs/DiagnosticReports/. Run the following to collect all Derp crash reports into a zip on your Desktop:

Terminal window
files=$(find ~/Library/Logs/DiagnosticReports -name "*.ips" -exec grep -l "dev\.derp" {} + 2>/dev/null) && [ -n "$files" ] && echo "$files" | xargs zip -j ~/Desktop/derp-crash-logs.zip || echo "No Derp crash reports found."

Attach the resulting derp-crash-logs.zip to your bug report.

Occasionally, the Derp team may ask you to provide debugging information on Windows OS in particular with one of the following:

Terminal window
# If Derp is in your PATH, Run:
derp --dump-debug-info
# Otherwise you may need to use an absolute path and ...
# Derp on Windows is installed for a single user, Run:
& $env:LOCALAPPDATA\programs\Derp\derp.exe --dump-debug-info
# Derp on Windows is installed for all users, Run:
& $env:PROGRAMFILES\Derp\derp.exe --dump-debug-info

Certain conditions can cause Derp to use more CPU than expected or become unresponsive. Collecting a CPU sample while the issue is happening is the best way to report it. The sample provides the information the Derp team needs to identify and fix the root cause.

Collect a sample using the steps for your platform and attach it to a new GitHub issue.

  1. Reproduce the high CPU usage or unresponsiveness in Derp.
  2. While the issue is occurring, open Activity Monitor, select the Derp process, and click Sample Process.

Sampling the Derp process in Activity Monitor

Sampling a Derp process in Activity Monitor.

  1. Save the resulting sample and attach it to your GitHub issue.

To gather the conversation ID, RIGHT-CLICK on the AI conversation block in question and select "Copy conversation ID", then paste that into the bug report that you submit so that our team can investigate the issue.

Whenever there is an error in the Agent Conversation, there will also be an option to directly copy the conversation ID for the bug report.

Agent Mode error message with Send Feedback button and debug information containing request and conversation IDs