beyhan/bifrost
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

first commit

Browse Source
This commit is contained in:
Beyhan Oğur
2026-04-26 21:52:23 +03:00
commit 880f412e2c
2662 changed files with 866266 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

102
docs/changelogs/v1.4.5.mdx Normal file
View File

@@ -0,0 +1,102 @@
---
title: "v1.4.5"
description: "v1.4.5 changelog - 2026-02-10"
---
<Tabs>
<Tab title="NPX">
```bash
npx -y @maximhq/bifrost --transport-version v1.4.5
```
</Tab>
<Tab title="Docker">
```bash
docker pull maximhq/bifrost:v1.4.5
docker run -p 8080:8080 maximhq/bifrost:v1.4.5
```
</Tab>
</Tabs>
<Update label="Bifrost(HTTP)" description="1.4.5">
- fix: duplicate mcp server creation when adding non oauth mcp client
- feat: add model and provider level governance - set budgets and rate limits on specific models or providers independent of virtual keys
- feat: cross-provider model matching - governance configs for `gpt-4o` now correctly apply to `openai/gpt-4o`, `gpt-4o-2024-08-06`, etc.
- feat: add `from_memory=true` query parameter for faster governance reads (virtual keys, model configs, provider governance)
- feat: added routing rules for dynamic routing of requests based on predefined rules
- feat: add `GET /api/models/base` endpoint for listing distinct base model names with search/filter support
- feat: base model selection in model limits UI when no provider is selected
- fix: edit sheets now show live data instead of stale cached values
- fix: mapping of multiple modality tokens from gemini usage metadata to bifrost usage
- fix: embedding thought signature in tool call id for valid tool calling cycle in gemini chat
- feat: request path override functionality to support full URLs (with scheme and host) as well as custom paths
- fix: missing and duplicated tool results in Bedrock - [@hhieuu](https://github.com/hhieuu)
- fix: errored request logs are now not counted in missing cost filter
- feat: adds support for custom OAuth scopes when authenticating with Azure Entra ID
- fix: if governance is disabled set enforce virtual key header to false
- feat: adds retry logic to MCP client connection establishment and tool retrieval
- fix: MCP server edit fixes for headers, ping available toggle and tool sync interval
- fix: force reload pricing now correctly resets the model pool and adds new models to the catalog
- feat: add routing engine used to log entries and telemetry metrics
- fix: tool sync interval in mcp catalog
</Update>
<Update label="Core" description="1.4.1">
- fix: model names with namespaces (e.g., `meta-llama/Llama-3.1-8B`) are now correctly preserved instead of being incorrectly split as provider-prefixed models
- fix: mapping of multiple modality tokens from gemini usage metadata to bifrost usage
- fix: embedding thought signature in tool call id for valid tool calling cycle in gemini chat
- feat: request path override functionality to support full URLs (with scheme and host) as well as custom paths
- fix: missing and duplicated tool results in Bedrock - [@hhieuu](https://github.com/hhieuu)
- fix: support HuggingFace model names without an explicit provider prefix
- feat: adds support for custom OAuth scopes when authenticating with Azure Entra ID
- feat: adds retry logic to MCP client connection establishment and tool retrieval
</Update>
<Update label="Framework" description="1.2.19">
- feat: add base_model support to model catalog for cross-provider model matching
- feat: add GetBaseModelName, IsSameModel, and GetDistinctBaseModelNames methods to ModelCatalog for resolving model aliases and checking model equivalence
- feat: add database migration for base_model column on model pricing table
- chore: upgrade core to 1.4.1
- feat: add database migration for routing_engine_used column on logs table
</Update>
<Update label="governance" description="1.4.20">
- feat: cross-provider model matching - budget/rate-limit configs for `gpt-4o` now apply to `openai/gpt-4o`, `gpt-4o-2024-08-06`, etc.
- feat: expand GovernanceData with ModelConfigs and Providers for in-memory reads
- feat: added routing rules for dynamic routing of requests based on predefined rules
- feat: add model and provider level governance - set budgets and rate limits on specific models or providers independent of virtual keys
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="jsonparser" description="1.4.19">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="litellmcompat" description="0.0.9">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="logging" description="1.4.19">
- feat: add routing engine used to log entries
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="maxim" description="1.5.18">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="mocker" description="1.4.19">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="otel" description="1.1.19">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="semantic_cache" description="1.4.18">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
</Update>
<Update label="telemetry" description="1.4.20">
- chore: upgrade core to 1.4.1 and framework to 1.2.19
- feat: add routing engine used to telemetry metrics
</Update>