IN THIS ARTICLE

AI Speech to text feature

Incorrect email!

The article was successfully sent to the email

Summary

Starting with TRBOnet 6.5, voice transmissions can be converted to text using AI language models. Two methods are supported: OpenAI commercial service and a self-hosted Local deployment. Transcription runs in real time or retrospectively in reports. Quality depends on the selected language model.

Do not install the local model on the same computer running TRBOnet. Local models impose heavy system load. Use a dedicated computer built for the model.

Speech to Text Processing

1. The server receives or sends a voice transmission.
2. The server removes silence segments from the transmission.
3. The server sends the prepared audio file to the language model using the endpoint /v1/audio/transcriptions.
4. The language model receives and processes the file.
5. The language model returns the transcribed text to TRBOnet.

Model Types

Method Model Endpoint
OpenAI API key Whisper-1 OpenAI servers
Local v1 compatible LLM model with audio support Self-hosted

OpenAI Model

Model: Whisper-1

Configuration: Enter the purchased API key in Administration System Settings AI Speech to Text. Requests are sent to OpenAI servers.

TRBOnet Demo and Trial license includes a pre-configured OpenAI Whisper test connection. This function is unavailable with Commercial license.

Local Model

Use any model compatible with v1 API. Deploy using publicly available solutions such as LocalAI.

Tested models: Whisper-1, Gemma-4.

Setup

1. Deploy the Local model (Example: https://github.com/mudler/LocalAI).
2. Add the language model.
3. Go to Administration System Settings AI Speech to Text.
4. In Model, enter the model name (example: Gemma-4).
5. In API endpoint, enter the Local Model server address: https://XXX.XXX.XXX.XXX/v1/audio/transcriptions.
6. In API key, enter the key configured in the Local model. If no API key is set up in the Local model, enter any text in this field.
7. In Language, select the language that system users will speak. This setting controls the language the model expects in audio input and uses for text output.

The language setting is required. Transcription accuracy depends on matching the selected language to the actual speech in transmissions.

For detailed configuration options and output example, see the following sections in the TRBOnet user manual:

6.4.5.15 AI Speech to Text - administration settings for the feature
6.5.8.3 Speech to Text - user-facing speech to text functionality
6.10 Voice Recording - voice recording and transcript integration

Recommendations

Use a language model suitable for the speaker language. Some models provide better transcription for a specific language.
Load models into GPU video memory (VRAM) for faster performance. Alternatively, use dedicated hardware solutions tailored for AI operation.
Test the local model under high load to determine acceptable performance.
Neocom Software does not provide hardware specifications. Requirements depend on the model and load level.

Helpful?
We're glad this article helped.

Thanks for letting us know. What went wrong?