Code base indexing returns 400

[springrider]

App Version

Version: 3.27.0 (173acdb)

API Provider

Google Gemini

Model Used

gemini

Roo Code Task Links (Optional)

No response

🔁 Steps to Reproduce

here is the settings:

💥 Outcome Summary

`I modified the code and logged the request & response info:

DEBUG: embeddings/user defined encoding_format:
base64
DEBUG: [log_20057d] sending request
{
  "method": "post",
  "url": "https://generativelanguage.googleapis.com/v1beta/openai/embeddings",
  "options": {
    "method": "post",
    "path": "/embeddings",
    "body": {
      "input": [
        "test"
      ],
      "model": "gemini-embedding-001",
      "encoding_format": "base64"
    }
  },
  "headers": {
    "accept": "application/json",
    "authorization": "***",
    "content-type": "application/json",
    "user-agent": "Pn/JS 5.12.2",
    "x-stainless-arch": "arm64",
    "x-stainless-lang": "js",
    "x-stainless-os": "MacOS",
    "x-stainless-package-version": "5.12.2",
    "x-stainless-retry-count": "0",
    "x-stainless-runtime": "node",
    "x-stainless-runtime-version": "v22.17.0"
  }
}
INFO: [log_20057d] post https://generativelanguage.googleapis.com/v1beta/openai/embeddings failed with status 400 in 554ms - error; not retryable
DEBUG: [log_20057d] response error (error; not retryable)
{
  "url": "https://generativelanguage.googleapis.com/v1beta/openai/embeddings",
  "status": 400,
  "headers": {
    "alt-svc": "h3=\":443\"; ma=2592000,h3-29=\":443\"; ma=2592000",
    "content-encoding": "gzip",
    "content-type": "application/json; charset=UTF-8",
    "date": "Sun, 07 Sep 2025 15:59:57 GMT",
    "server": "scaffolding on HTTPServer2",
    "server-timing": "gfet4t7; dur=230",
    "transfer-encoding": "chunked",
    "vary": "Origin, X-Origin, Referer",
    "x-content-type-options": "nosniff",
    "x-frame-options": "SAMEORIGIN",
    "x-xss-protection": "0"
  },
  "durationMs": 555
}

I did some research that look like the parameter is wrong? it should not used openAI's format to call gemini's API? `

📄 Relevant Logs or Errors (Optional)

[roomote]

I'm exploring a potential fix for this issue. The problem appears to be that the encoding_format parameter (which is OpenAI-specific) is being incorrectly sent to Google Gemini's embeddings API, causing a 400 error. I'll work on ensuring the API request format is compatible with Gemini's requirements.

[roomote]

I've opened PR #7762 with a fix for this issue: #7762

The fix detects Google Gemini endpoints and excludes the OpenAI-specific encoding_format parameter that was causing the 400 error. All CI checks have passed successfully.

The implementation:

• Adds detection for Gemini URLs to conditionally exclude the parameter
• Maintains backward compatibility for all other OpenAI-compatible endpoints
• Includes comprehensive test coverage

This should resolve the code base indexing issue when using Google Gemini as the embeddings provider.

[daniel-lxs]

Hey, I mainly use Gemini for embeddings and I'm not seeing this issue at all, the endpoint used is the openAI compatible endpoint that Gemini offers, should work normally and at least in my case it is. Can you reenter all your settings again and see if that works?