Skip to content

Add Exmaples for how to use A2UI on Gemini Enterprise#1043

Merged
yuantian-agentspace merged 8 commits intomainfrom
gemini_enterprise
Apr 6, 2026
Merged

Add Exmaples for how to use A2UI on Gemini Enterprise#1043
yuantian-agentspace merged 8 commits intomainfrom
gemini_enterprise

Conversation

@yuantian-agentspace
Copy link
Copy Markdown
Collaborator

Description

Provide examples of how to use A2UI on Gemini Enterprise. It covers Cloud Run deployment and Agent Engine Deployment.

Pre-launch Checklist

If you need help, consider asking for advice on the discussion board.

Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request introduces comprehensive samples and scripts for deploying A2UI agents on Gemini Enterprise using both Vertex AI Agent Engine and Cloud Run. The implementation includes a base agent executor with A2UI validation, sample UI templates, and registration logic. The review feedback identifies several critical issues that need to be addressed: a shell syntax error in the documentation's curl command, potential data corruption in the JSON cleaning logic due to incorrect string stripping, the use of invalid Gemini model identifiers across multiple files, and a dependency mismatch in the requirements file that would result in runtime import errors.

@wrenj
Copy link
Copy Markdown
Collaborator

wrenj commented Apr 1, 2026

Hi Yuan, Thanks for the PR.

Can you work with @nan-yu to use the python SDK he built for things like loading samples, loading the catalog, generating system instructions, etc? It'll let this sample focus on the interesting behavior (GE integration) and stop using some antipatterns we have in our earlier samples like copy and pasting the catalog json into a python string.

Copy link
Copy Markdown
Collaborator

@nan-yu nan-yu left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/LGTM

"--- ContactAgent.fetch_response: Received no final response"
f" content from runner (Attempt {attempt}). ---"
)
if attempt <= max_retries:
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

im not sure we need this retries TBH, we already auto correct malformed json and it complicates the code, but feel free to leave if it you find them useful @nan-yu

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it's okay to keep the retry here as the auto-fixer only removes trailing commas for now. There're other JSON failures, such as fields in snake_case vs. camelCase.

The retry may not be triggered because the examples have the correct format, but it may serve as a guardrail for other random failures.

@yuantian-agentspace yuantian-agentspace merged commit a3416be into main Apr 6, 2026
6 checks passed
@github-project-automation github-project-automation bot moved this from Todo to Done in A2UI Apr 6, 2026
@yuantian-agentspace yuantian-agentspace deleted the gemini_enterprise branch April 6, 2026 21:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

3 participants