--- title: "v1.4.8" description: "v1.4.8 changelog - 2026-02-25" --- ```bash npx -y @maximhq/bifrost --transport-version v1.4.8 ``` ```bash docker pull maximhq/bifrost:v1.4.8 docker run -p 8080:8080 maximhq/bifrost:v1.4.8 ``` - feat: added support for logging headers to capture request headers into log metadata - fix: OAuth callback URL now respects X-Forwarded-Proto header for correct HTTPS scheme behind reverse proxies - feat: add asynchronous inference support - feat: added required headers support — configure headers that must be present on every request - fix: routing rules CEL builder UI fixes - fix: routing rules now support case-insensitive header matching - fix: azure openai sdk requests handling in openai integration - feat: support multiple credential types in vertex auth credentials - fix: semantic caching plugin initialization when configured using UI - fix: added replicate and huggingface model allowlist on list models response - fix: added support for unfiltered list models response when updating keys' allowed models field - fix: backfill allowed models that were not in the list models response - feat: support for video generation requests - feat: add runway provider support - enhancement: improved model matching logic in model catalog - feat: support for preserving JSON key order in tool function parameters - chore: bump core version to 1.4.4 and framework to 1.2.22 - refactor: ListModelsRequest to use the common request handling pipeline instead of its own implementation - feat: added support for filtering /v1/models responses based on virtual key configurations in the governance plugin - feat: add vllm provider support - feat: support multiple credential types in vertex auth credentials - fix: added replicate and huggingface model allowlist on list models response - feat: added support for unfiltered list models response - fix: backfill allowed models that were not in the list models response - feat: support for video generation requests - feat: add runway provider support - feat: support for preserving JSON key order in tool function parameters - fix: empty json marshalling fixes for tool choice struct - feat: added logging_headers config field and metadata column to MCP tool logs - feat: add tables for async job results - feat: added required_headers config field with DB persistence and migration - feat: added support for unfiltered models for provider in model catalog - enhancement: improved model matching logic in model catalog - chore: bump core version to 1.4.4 - fix: routing rules now support case-insensitive header matching - feat: added required headers validation — requests missing configured headers are rejected with a 400 error - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22 - feat: added support for capturing configured logging headers and x-bf-lh-* prefixed headers into log metadata for both LLM and MCP requests - feat: add metadata column to logging table for async request tracking - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22 - fix: invalid providers which do not support embedding requests are rejected during initialization - chore: bump core version to 1.4.4 and framework to 1.2.22 - chore: bump core version to 1.4.4 and framework to 1.2.22