Moderation
Moderation
Content moderation refers to the review of user-inputted content and content generated by agents for elements such as pornography, violence, etc., ensuring content health and safety, and reducing clients' business compliance risks.
Service
Currently, OpenAI Moderation is supported.
Key Configuration
- GPTBots Key: This allows users to directly use the corresponding content moderation service via the Key provided by GPTBots. There is no need for users to separately integrate with OpenAI.
- My Key: Users can also use their own OpenAI Key to access the corresponding content moderation service.
API Extension
If you need to use content moderation services from other providers, you can refer to the custom API extension method below for integration and expansion.
API Development
To use a custom API for content moderation, the API must be developed in accordance with the following requirements:
- It must support
Bearer
authentication. - It must use the
POST
method. - The response body must follow the structure below:
{
"flagged": true,
"preset_response": "Your content violates our usage policy."
}
{
"flagged": true,
"preset_response": "Your content violates our usage policy."
}
このコードブロックは、フローティングウィンドウに表示されます
Field | Type | Description |
---|---|---|
flagged | Bool | Indicates the moderation result, whether the content has passed the review. true means not passed, and false means passed. |
preset_response | String | When the content moderation result is not passed, this text will be used to reply to the user. |
Add API
Click "New API Extension"。
Configure the information of the extended API into the panel, and click "Test".
If the test is successful, click "Save", and the API extension will then take effect.