Speech-to-Text (ASR) Capability Integration

← Developer Docs · 简体中文

Is Code Required?

Always required. The ASR API has no industry-standard format. Each vendor uses different proprietary API formats, so regardless of whether the platform is OpenAI-compatible for chat, you need to write a Service class.

Note: The ASR service is specified by system-level configuration; users cannot switch between providers. Only one ASR service is active at a time. The system determines the active ASR service via the adi_sys_config table record where name = 'asr_setting'. The value is in JSON format:

{"model_name": "new-model-name", "platform": "newai", "max_record_duration": 60, "max_file_size": 10485760}

You can change this via the admin backend, or update the database directly:

UPDATE adi_sys_config SET value = '{"model_name":"newai-asr-v1","platform":"newai","max_record_duration":60,"max_file_size":10485760}' WHERE name = 'asr_setting';

Prerequisite: Register Platform Name Constant

Add a new platform name constant in AdiConstant.ModelPlatform:

File: adi-common/src/main/java/com/moyz/adi/common/cosntant/AdiConstant.java

public static class ModelPlatform {
    // ... existing constants ...
    public static final String NEWAI = "newai";
}

Step 1: Service Development

Inheritance: CommonModelService → AbstractAsrModelService → Your Service

The base class AbstractAsrModelService is thin (~20 lines), defining only the audioToText abstract method and proxy settings, so most of the logic must be implemented by subclasses.

Required abstract methods:

Method	Return Type	Description
audioToText(String urlOrPath)	String	Receive an audio file URL or local path, return recognized text

Complexity Notes

Implementation complexity is moderate. It requires handling file uploads (local or remote URLs), HTTP requests, response parsing, etc.:

• Async API mode (see DashScopeAsrService, ~76 lines): Uses SDK to submit an async task, polls for results. Supports remote URLs only
• Sync upload mode (see SiliconflowAsrService, ~90 lines): Uses multipart/form-data to upload audio files, returns results synchronously. Supports both local files and remote URLs

Step 2: Register in AiModelInitializer

File: adi-common/src/main/java/com/moyz/adi/common/service/AiModelInitializer.java

Add in the initAsrModelServiceList() method:

private synchronized void initAsrModelServiceList(Map<String, ModelPlatform> nameToPlatform) {
    // ... existing platforms ...

    // Add new
    initAsrModelService(AdiConstant.ModelPlatform.NEWAI,
        model -> new NewAiAsrService(model, nameToPlatform.get(AdiConstant.ModelPlatform.NEWAI)));
}

---

Previous: TTS Integration · Next: Back to contents