SiliconFlow Integration

This page explains how to connect Bytedesk to SiliconFlow models and use SiliconFlow as a cloud inference provider.

Prerequisites

• Bytedesk has been deployed
• You have created a SiliconFlow API key

Configuration Steps

1. Create an API key in SiliconFlow: https://www.siliconflow.cn
2. Sign in to Bytedesk admin and enter the key
3. Select SiliconFlow as the default provider
4. Generate and embed the chat code

Example Result

After setup, Bytedesk can invoke SiliconFlow-backed model services.

Optional Configuration

SPRING_AI_SILICONFLOW_BASE_URL=https://api.siliconflow.cn
SPRING_AI_SILICONFLOW_API_KEY=sk-xxxx
SPRING_AI_SILICONFLOW_CHAT_ENABLED=true
SPRING_AI_SILICONFLOW_CHAT_OPTIONS_MODEL=Qwen/QwQ-32B
SPRING_AI_SILICONFLOW_CHAT_OPTIONS_TEMPERATURE=0.7

spring.ai.siliconflow.base-url=https://api.siliconflow.cn
spring.ai.siliconflow.api-key=sk-xxxx
spring.ai.siliconflow.chat.enabled=true
spring.ai.siliconflow.chat.options.model=Qwen/QwQ-32B
spring.ai.siliconflow.chat.options.temperature=0.7

Common Issues

1. Invalid key: confirm the SiliconFlow key is enabled.
2. Slow responses: inspect network and provider load.
3. Feature errors: verify the selected model is available.

Related Resources

• SiliconFlow Documentation
• Spring AI SiliconFlow
• Bytedesk Docs