Making HTTP Requests More Agile & Adaptive

[ForeverYoungFly]

Hello, thank you for your outstanding work first and foremost.

However, I've encountered an issue or rather, a suggestion: would it be possible to incorporate an HTTP request model?

I've built an AI application using the qwen-max model, which embeds my knowledge base. Although custom models were added in version 2.6.0, under this model, I'm unable to invoke a specific AI model application for service via the app_id. Have you considered adding the functionality to invoke models via HTTP POST in the future? This way, users could leverage HTTP to call AI applications they've created within third-party models.

Regardless, your work is highly commendable, and I express my gratitude once again!

[logancyang]

@ForeverYoungFly what do you mean by http request? It is using HTTP request.

• The plugin supports any 3rd party models that has openai format. Does your model have openai format?
• Did it work before v2.6.0?