mirror of
https://github.com/langgenius/dify-official-plugins.git
synced 2026-07-22 01:55:27 -04:00
Error with Vertex AI google models #801
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @tahasuperr on GitHub (Nov 13, 2025).
Self Checks
Dify version
Version 1.9.2
Cloud or Self Hosted
Cloud
Steps to reproduce
Run failed: req_id: 1892037c54 PluginInvokeError: {"args":{"description":"[models] Error: 'NoneType' object has no attribute 'get'"},"error_type":"InvokeError","message":"[models] Error: 'NoneType' object has no attribute 'get'"}
Getting this error in all my nodes with google vertex api, it is working with Open AI
✔️ Expected Behavior
It should work
❌ Actual Behavior
Run failed: req_id: 1892037c54 PluginInvokeError: {"args":{"description":"[models] Error: 'NoneType' object has no attribute 'get'"},"error_type":"InvokeError","message":"[models] Error: 'NoneType' object has no attribute 'get'"}
@dosubot[bot] commented on GitHub (Nov 13, 2025):
This error usually means that a required configuration for the Vertex AI model is missing or not structured as expected, causing the code to call
.geton aNoneobject. Compared to OpenAI, the Vertex AI integration in Dify is less robust against missing configs, so if any required field (likemodel_parameters) is absent or misconfigured, you'll see this error.Make sure all necessary fields for your Vertex AI provider are present and correctly structured in your configuration. Double-check your model provider settings in Dify, especially for fields like project ID, region, model name, and authentication. If you are using a plugin, ensure it's up to date and compatible with your Dify version.
If you recently updated the Vertex AI plugin or Dify itself, some users have found that downgrading the VertexAI plugin to an earlier version (such as 0.0.6) or updating to the latest available version can resolve similar issues. Also, some errors have been fixed in later Dify and plugin releases, so upgrading both Dify and the VertexAI plugin may help if you are not already on the latest versions. See more details and user experiences in this related thread: VertexAI plugin issues and fixes.
If the problem persists, check your plugin daemon logs for clues about missing or misconfigured parameters, and ensure your environment variables and credentials are set up as required by Vertex AI.
To reply, just mention @dosu.
How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other
@dosubot[bot] commented on GitHub (Nov 13, 2025):
This error is a known issue in Dify v1.9.2 when using Google Vertex AI models, especially Gemini, and is caused by a mismatch between the expected response format and what Vertex AI returns—often when Dify uses the
:predictendpoint instead of the required:streamGenerateContentendpoint for Gemini models. The error happens when the plugin returnsNoneor a malformed response, and Dify tries to access.geton it, which isn't robustly handled in the current codebase. This doesn't happen with OpenAI because its plugin always returns the expected data structure reference.Workarounds reported by users include: using supported regions like
us-central1, switching to non-preview Gemini models, retrying requests if errors are intermittent, or using Gemini 2.5 flash-lite for more stability reference. Also, double-check that your Vertex AI credentials and region are correctly configured—setting region toglobalis not supported for embeddings, for example reference.A migration to the new
google-genaiSDK with better response handling has been implemented in a recent PR, but it may not be present in v1.9.2 cloud reference. Until the plugin is updated, you may need to use the workarounds above or monitor for future updates that add defensive checks and proper endpoint support.If you control any workflow or plugin code, adding defensive checks for missing keys (using
.get()or fallback values) can help avoid these errors reference.To reply, just mention @dosu.
How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other
@nadazero commented on GitHub (Nov 13, 2025):
I met same error.
@Allamss commented on GitHub (Nov 14, 2025):
This issue can be reproduced in both versions 0.0.26 and 0.0.27, but it works again after downgrading to 0.0.25.
@nadazero commented on GitHub (Nov 19, 2025):
@Zhaohuai-L please have a look at this issue and fix this bug ASAP please, this issue will block testing new model such as gemini 3.0
@dosubot[bot] commented on GitHub (Dec 6, 2025):
Hi, @tahasuperr. I'm Dosu, and I'm helping the dify-official-plugins team manage their backlog and am marking this issue as stale.
Issue Summary:
What's Next:
Thank you for your understanding and contribution!