maximhq/bifrost transports/v1.1.15

on GitHub

What's Changed

• chore: update bifrost/core to v1.1.13 and remove local replace by @Pratham-Mishra04 in #205
• feat: add embeddings support to HTTP transport by @Pratham-Mishra04 in #207
• feat: add embedding logging ui by @Pratham-Mishra04 in #208
• feat: add embeddings support to integration tests by @Pratham-Mishra04 in #212
• refactor: params shifted to root level for completion request by @TejasGhatte in #200

Sure! Here's a clearer and more polished version of your note:

📢 Note on Bifrost HTTP Transport Update

In this release, we've simplified how model parameters are sent in the request body when using Bifrost HTTP transport.

You no longer need to wrap model parameters inside a params field. Parameters can now be passed directly at the top level of the request body.

🔄 Before (Old Format):

{
  "model": "openai/gpt-4o",
  "params": {
    "max_token": 1024
  }
}

✅ Now (New Format):

{
  "model": "openai/gpt-4o",
  "max_token": 1024
}

This change makes requests cleaner and more consistent.

Full Changelog: core/v1.1.13...transports/v1.1.15