The same API works in "continue" but doesn't work in "roo code."

[CY-W]

Which version of the app are you using?

v3.3.23

Which API Provider are you using?

OpenAI Compatible

Which Model are you using?

deepseek by company deployment

What happened?

In the "continue" environment, making the following edits allows the API to be used normally; however, such configuration in the "roo code" is not feasible. How can I configure the API provider in the "roo code" through a configuration file, just like in the "continue" environment?

Steps to reproduce

As shown in the figure above, the company-provided DeepSeek API was configured in both "continue" and "roo code." The API works normally in "continue," but it has always been in the request state in "roo code."

I suspect that the issue lies in the "provider" field in the configuration file. In "roo," it is set to "openai compatible," while in "continue," it is set to "openai." Could you please advise on how to directly edit the configuration file of "roo code"?

Relevant API REQUEST output

Additional context

【update】
The current deployment uses the third-party Deepseek model. The API provider indicates they have not yet adapted to the new OpenAI API specifications. Implementation steps (as shown in the diagram above) involved only inputting the necessary API parameters, but encountered errors when submitting queries. However, under the similar configuration, Alibaba's Qwen model functions properly. Does version 3.7.12 employ the updated OpenAI API specifications?

[dosubot]

Hey @CY-W! I'm here to help you with any bugs, questions, or contributions you might have. Let's get to the bottom of this together!

To configure the API provider in "Roo Code" similarly to how it's done in "Continue," you need to ensure that the settings in the "Roo Code" configuration file are correctly specified for OpenAI-compatible providers. Here are the steps you can follow:

1. Configuration File: Ensure that the configuration file for "Roo Code" includes the correct fields for the API provider. You should specify the Base URL, API Key, and Model ID. These settings are typically found in the Roo Code settings panel.

2. Provider Field: The issue might be related to the "provider" field. In "Roo Code," it should be set to "openai-compatible" if you are using an OpenAI-compatible provider. Double-check that this matches the expected format for "Roo Code."

3. Token Length Error: The 422 status code error indicates a problem with the token length. Ensure that the messages being sent do not exceed the token limits specified by the API. This might require adjusting the input or payload size.

4. Documentation: For detailed guidance on setting up OpenAI-compatible providers in "Roo Code," refer to the documentation section titled "Using OpenAI Compatible Providers With Roo Code" available at https://docs.roocode.com/providers/openai-compatible ^[1].

By ensuring these configurations are correctly set, you should be able to resolve the issue with the API provider in "Roo Code."

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other}

[hannesrudolph]

Thanks for the report! Can you please test is the latest version 3.7.12 has solved this issue? If not, would you be able to provide clear steps to reproduce the issue?

[CY-W]

Unfortunately, after updating to the latest version, the issue remains unresolved, with the problem manifesting in the same way.
Is it because Roo Code uses the new version of the OpenAI interface, while the Continue plugin uses the old version?

[hannesrudolph]

would you be able to provide clear steps to reproduce the issue?

[hannesrudolph]

Will reopen it if you can either help me reproduce the error or make it clear what the problem is. I am less concerened about what Continue does and more concerned about exactly what is happening and when.