# Greptile
**URL:** https://discourse.itk.org/t/greptile/7737
**Category:** Engineering
**Tags:** itk
**Created:** [March 23, 2026, 5:49pm UTC](https://discourse.itk.org/t/greptile/7737 "2026-03-23T17:49:14Z")
**Posts on this page:** 7
**Page:** 1
### Author:  [@dzenanz](https://discourse.itk.org/u/dzenanz)
#### Post date: [March 23, 2026, 5:49pm UTC](https://discourse.itk.org/t/greptile/7737/1 "2026-03-23T17:49:14Z")
I got an email from Greptile:
| Greptile lets you add custom rules to enforce your organization’s preferences, coding standards and best practices automatically across pull requests.
You can configure them in two ways:
|
| --- |
| |
1. [.greptile/ Configuration](https://c.vialoops.com/CL0/https:%2F%2Fwww.greptile.com%2Fdocs%2Fcode-review%2Fgreptile-config/1/0100019d0cc7b178-efbd2bcc-8e23-4dbe-9583-c5ec3dad5a60-000000/-S4x_Jd0kAqchjwrdtD9ADjzlSlEdYgZHnqKooXW4Ok=449): Individual teams can tailor settings, rules and context at the directory level in a .greptile/ folder.
| |
| --- |
1. [Greptile Dashboard](https://c.vialoops.com/CL0/https:%2F%2Fwww.greptile.com%2Fdocs%2Fcode-review%2Fcustom-standards%23method-1-dashboard/1/0100019d0cc7b178-efbd2bcc-8e23-4dbe-9583-c5ec3dad5a60-000000/P01jbYA_94bIiYmlUyFOT_Ma6QiP7K0fhN0HwAIHlhw=449): For quick experiments, or org wide defaults, you can add rules directly to the dashboard.
|
These patterns are optional and not required for high quality code reviews, but are recommended for orgs to create a tailored experience. |
| --- |
If someone wants to customize our Greptile bot, it is possible.
---
### Author:  [@blowekamp](https://discourse.itk.org/u/blowekamp)
#### Post date: [March 25, 2026, 1:17pm UTC](https://discourse.itk.org/t/greptile/7737/2 "2026-03-25T13:17:28Z")
I am finding the flow charts are not useful enough to justify their size when scrolling.
---
### Author:  [@dzenanz](https://discourse.itk.org/u/dzenanz)
#### Post date: [March 25, 2026, 1:31pm UTC](https://discourse.itk.org/t/greptile/7737/3 "2026-03-25T13:31:43Z")
I made some setting changes through the [web UI](https://app.greptile.com/review/github). I don’t know how quickly will these settings propagate. I also invited you and a few others as admins of Greptile organization.
---
### Author:  [@Niels\_Dekker](https://discourse.itk.org/u/Niels_Dekker)
#### Post date: [March 26, 2026, 10:55am UTC](https://discourse.itk.org/t/greptile/7737/4 "2026-03-26T10:55:15Z")
> [@blowekamp](#):
>
> I am finding the flow charts are not useful enough to justify their size when scrolling.
I very much agree! In general the output of Greptile appears too elaborate to me, especially for small pull requests. For example, the following PR just removed 9 lines of code, and added one line:
[MarkovRandomFieldsClassifiers: Replace buffered region iteration in MRFImageFilter, remove iterator from RGBGibbsPriorFilter by N-Dekker · Pull Request #5977 · InsightSoftwareConsortium/ITK · GitHub](https://github.com/InsightSoftwareConsortium/ITK/pull/5977)
Still it got quite extensive feedback from Greptile!
More in general, the question is how we should deal with AI generated reviews. As a submitter of a PR, I hope that human reviewers keep focused on the actual code changes that I’m proposing, as well as the proposed commit messages. I hope that reviewers don’t get distracted by the amount of Greptile output.
---
### Author:  [@dzenanz](https://discourse.itk.org/u/dzenanz)
#### Post date: [March 26, 2026, 12:37pm UTC](https://discourse.itk.org/t/greptile/7737/5 "2026-03-26T12:37:51Z")
>
>
> \- Following pull request #5839
This PR has the long parts of Greptile reviews collapsed by default. So the changes I made to settings yesterday are having an effect.
---
### Author:  [@Niels\_Dekker](https://discourse.itk.org/u/Niels_Dekker)
#### Post date: [March 26, 2026, 12:55pm UTC](https://discourse.itk.org/t/greptile/7737/6 "2026-03-26T12:55:58Z")
Thanks @dzenanz
It’s good to be aware of the fact that AI reviews may themselves have flaws. I just noticed one in the flowchart generated by Greptile: [Remove local type aliases of ImageRegionIterator and ImageRegionConstIterator in `*.hxx` files by N-Dekker · Pull Request #5981 · InsightSoftwareConsortium/ITK · GitHub](https://github.com/InsightSoftwareConsortium/ITK/pull/5981#issuecomment-4134465943)
Also it appears to make mistakes when counting the number of (eliminated) variables of a pull request: [Replace buffered region iteration with FillBuffer in FastMarching filters by N-Dekker · Pull Request #5970 · InsightSoftwareConsortium/ITK · GitHub](https://github.com/InsightSoftwareConsortium/ITK/pull/5970#discussion_r2969536135)
Should we report those errors to the makers of Greptile? At least we should be aware of such flaws.
---
### Author:  [@dzenanz](https://discourse.itk.org/u/dzenanz)
#### Post date: [March 26, 2026, 1:38pm UTC](https://discourse.itk.org/t/greptile/7737/7 "2026-03-26T13:38:26Z")
> [@Niels\_Dekker](#):
>
> Should we report those errors to the makers of Greptile?
Give it a try, and see whether they are appreciative of the report.