feat(vertex_ai): multi-region Vertex hosts (aiplatform.*.rep.googleapis.com)

[milan-berri]

Relevant issues

Fixes #25926

Reference: Multi-region endpoints for Claude on Vertex AI (Google Cloud Blog)

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• I have Added testing in the tests/test_litellm/ directory, Adding at least 1 test is a hard requirement - see details
• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem
• I have requested a Greptile review by commenting @greptileai and received a Confidence Score of at least 4/5 before requesting a maintainer review

CI (LiteLLM team)

CI status guideline:

• 50-55 passing tests: main is stable with minor issues.
• 45-49 passing tests: acceptable but needs attention
• <= 40 passing tests: unstable; be careful with your merges and assess the risk.

• Branch creation CI run
  Link:

• CI run for the last commit
  Link:

• Merge / cherry-pick CI run
  Links:

Type

🆕 New Feature
✅ Test

Changes

• litellm/llms/vertex_ai/common_utils.py — get_vertex_base_url: After global, treat locations that are truthy and contain no - as Vertex multi-region hosts (https://aiplatform.{geo}.rep.googleapis.com). Otherwise keep the regional pattern https://{region}-aiplatform.googleapis.com (preserves prior behavior for None / empty via the final branch).
• litellm/proxy/pass_through_endpoints/llm_passthrough_endpoints.py: Local get_vertex_base_url duplicates the same rules with trailing slashes for URL joining (no import from common_utils). WebSocket upstream host for Vertex live passthrough uses the same multi-region vs regional logic so us / eu work for wss://.
• tests/test_litellm/llms/vertex_ai/test_vertex_global_url_support.py: Parametrize us and eu expected base URLs for get_vertex_base_url.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Apr 17, 2026 3:08pm

[codecov]

Codecov Report

❌ Patch coverage is 83.33333% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...ass_through_endpoints/llm_passthrough_endpoints.py	66.66%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[greptile-apps]

Greptile Summary

This PR adds support for Vertex AI multi-region endpoints (aiplatform.{geo}.rep.googleapis.com) by extending get_vertex_base_url in both common_utils.py and the passthrough module. The detection heuristic — a non-empty location string containing no - (and not equal to "global") — correctly distinguishes multi-region geographies (e.g. us, eu) from regional locations (e.g. us-central1). The previous review concerns around duplicate WebSocket logic and missing passthrough tests have both been resolved in this revision.

Confidence Score: 5/5

Safe to merge — changes are minimal, backward-compatible, and well-tested with mocked unit tests covering all location types.

All findings from the previous review round (duplicated WebSocket ternary, missing passthrough tests) have been addressed. The multi-region heuristic ("-" not in vertex_location, excluding global) is sound for current GCP region naming conventions. No regressions to None, empty string, or regional locations. All P2 or no-issue.

No files require special attention.

Important Files Changed

Filename	Overview
litellm/llms/vertex_ai/common_utils.py	Adds multi-region branch ("-" not in vertex_location) to get_vertex_base_url; preserves prior behavior for None, empty string, and regional locations.
litellm/proxy/pass_through_endpoints/llm_passthrough_endpoints.py	Mirrors multi-region logic in module-local get_vertex_base_url (with trailing slash); WebSocket handler now delegates to the helper instead of repeating inline ternary.
tests/test_litellm/llms/vertex_ai/test_vertex_global_url_support.py	Adds us and eu parametrized cases to TestVertexBaseURL; no real network calls, consistent with existing test patterns.
tests/test_litellm/proxy/pass_through_endpoints/test_llm_pass_through_endpoints.py	New TestVertexPassthroughGetVertexBaseUrl class covers both trailing-slash and WebSocket host-strip behavior for all four location types.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[vertex_location input] --> B{== 'global'?}
    B -->|Yes| C["https://aiplatform.googleapis.com[/]"]
    B -->|No| D{truthy AND no '-'?}
    D -->|Yes - multi-region e.g. us, eu| E["https://aiplatform.{geo}.rep.googleapis.com[/]"]
    D -->|No - regional or None| F["https://{location}-aiplatform.googleapis.com[/]"]
    C --> G[WebSocket: strip https:// + trailing /]
    E --> G
    F --> G
    G --> H["wss://{host}/ws/...BidiGenerateContent"]

Loading

_{Reviews (3): Last reviewed commit: "fix(vertex_ai): use aiplatform.{geo}.rep..." | Re-trigger Greptile}

[milan-berri]

@greptileai - review again

[jamesseiwert]

@milan-berri thank you for this! It is currently a blocker for us getting on opus-4-7. Looking forward to testing this

[mateo-berri]

Can you run uv lock and commit the new lock file? Seems the tests are failing

[milan-berri]

re-created to run CI proprely here #26281