Uploaded December 2025 | Updated September 2026, 2 weeks ago
In this video, Julian and Isaac from H2O.ai walk through the h2oGPTe GitHub Action, a simple way to bring AI agents into your GitHub workflows for PR reviews, debugging, and research on your infrastructure.
Try it yourself:
GitHub repo: h2oGPTe Action
What you'll see:
Quick install: Set up with Premium or a custom server, add API Key/API Base secrets, commit the workflow, and push.
PR review: Tag @h2oGPTe in a pull request comment to get strengths, gaps, and actionable next steps.
Debugging from issues: Post an error screenshot in an issue and ask @h2oGPTe to fix—it can create a PR, commit changes, and pass tests.
Auto review agent: Drop a ready-to-use h2oGPTe_autopr.yaml into .github/workflows to auto-review new PRs.
Deep research: Ask @h2oGPTe to recommend packages and usage (e.g., Fastify Caching, Node Cache) for features you're building.
Chapters:
00:00 Intro
00:54 Install walkthrough (API secrets + workflow)
03:25 PR review via comment
04:43 Debugging issues with screenshots
06:18 Custom prompts & auto PR review agent
07:26 Deep research for request caching
08:11 Wrap-up
Who it's for:
Developers who want AI assistance directly in GitHub
Teams looking to standardize PR reviews and speed up debugging
Anyone exploring agent workflows for software development
Questions or feedback? Drop a comment and we'll help you get set up.
In this video, Julian and Isaac from H2O.ai walk through the h2oGPTe GitHub Action, a simple way to bring AI agents into your GitHub workflows for PR reviews, debugging, and research on your infrastructure.
Try it yourself:
GitHub repo: h2oGPTe Action
What you'll see:
Quick install: Set up with Premium or a custom server, add API Key/API Base secrets, commit the workflow, and push.
PR review: Tag @h2oGPTe in a pull request comment to get strengths, gaps, and actionable next steps.
Debugging from issues: Post an error screenshot in an issue and ask @h2oGPTe to fix—it can create a PR, commit changes, and pass tests.
Auto review agent: Drop a ready-to-use h2oGPTe_autopr.yaml into .github/workflows to auto-review new PRs.
Deep research: Ask @h2oGPTe to recommend packages and usage (e.g., Fastify Caching, Node Cache) for features you're building.
Chapters:
00:00 Intro
00:54 Install walkthrough (API secrets + workflow)
03:25 PR review via comment
04:43 Debugging issues with screenshots
06:18 Custom prompts & auto PR review agent
07:26 Deep research for request caching
08:11 Wrap-up
Who it's for:
Developers who want AI assistance directly in GitHub
Teams looking to standardize PR reviews and speed up debugging
Anyone exploring agent workflows for software development
Questions or feedback? Drop a comment and we'll help you get set up.
