Audio transcription does not show up on gui after Audio transcription completed for file

Describe the bug
when ai post processing is enabled, audio transcription does not show up on gui after audio transcription completed

transcript text should be stored in a special kind of div/container, which is not edited by ai

Screenshots

Audio transcription completed for file: /app/.blinko/files/my_recording_176221212121_17612121213.webm,hello world。
Added transcriptions to note 16,world。

Desktop (please complete the following information):

• OS: Linux
• Browser chrome,
• Version . 22

Additional context
fixing also #894

[dosubot]

Hi @Busboy3129! I'm Dosu and I’m helping the blinko team.

The root cause is that the audio transcription logic for the 'openai' provider does not apply a custom baseURL, even if one is set in your settings. The code always defaults to https://api.openai.com/v1 for OpenAI, while the 'custom' provider path does respect the custom endpoint. To fix this, update the provider logic so that if a baseURL is set in the config, it is always passed to the OpenAI client—regardless of provider type. This will ensure audio transcription requests use your custom endpoint as expected and resolve the 401 errors when using non-default endpoints. You can see this difference in how the AudioProvider handles 'openai' vs 'custom' providers here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[cx88982060]

也就是说只能使用OpenAI提供的API才行？我用的硅基流动也不会自动语音转文字

---

In other words, you can only use the API provided by OpenAI? The silicon-based mobile phone I use cannot automatically convert speech to text

[ZenoBell]

Still experiencing this bug on the latest version.

Setup: Self-hosted Blinko on Docker, using fedirz/faster-whisper-server as a local Whisper endpoint (http://192.168.10.x:9001/v1).

The model list is fetched correctly, transcription completes successfully on the server side (logs show Audio transcription completed for file), but the transcribed text never appears in the UI.

Confirmed the request never hits the custom Whisper endpoint , the container logs show zero incoming requests when recording from Blinko, confirming it's still routing to api.openai.com/v1 instead of the configured custom URL.

Would love to see this fixed , the rest of the AI stack works perfectly with local Ollama. Voice transcription is the last missing piece for a fully local/offline setup. Any ETA or workaround in the meantime?