> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getbifrost.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# v1.5.0-prerelease6

> v1.5.0-prerelease6 changelog - 2026-04-24

<Tabs>
  <Tab title="NPX">
    ```bash theme={null}
    npx -y @maximhq/bifrost --transport-version v1.5.0-prerelease6
    ```
  </Tab>

  <Tab title="Docker">
    ```bash theme={null}
    docker pull maximhq/bifrost:v1.5.0-prerelease6
    docker run -p 8080:8080 maximhq/bifrost:v1.5.0-prerelease6
    ```
  </Tab>
</Tabs>

<Update label="Bifrost(HTTP)" description="1.5.0-prerelease6">
  ## ✨ Features

  * **MCP Tool Groups** - Added `tool_groups` config with governance scoping (virtual key, team, customer, user, provider, API key) and camelCase Helm aliases for MCP client fields
  * **Enterprise Helm Overlays** - Suite of composable overlays for guardrails, org governance, access profiles, customer budgets, teams, multi-customer governance, and SCIM/SSO
  * **Semantic Cache Helm Layers** - Added `values-semantic-search-redis.yaml` and `values-semantic-search-weaviate.yaml` for Redis and Weaviate-backed semantic caches, plus a client-config overlay
  * **Period Parameter in Dashboard/Logs APIs** - Added `period` param for relative time range queries on dashboard and logs endpoints
  * **`provider_key_name` Alias** - Human-readable alias for routing targets and pricing overrides, resolved to `key_id` at config load time
  * **`env.*` References for Proxy and TLS** - Proxy and TLS config fields (`url`, `username`, `password`, `ca_cert_pem`) now accept `env.VAR_NAME` for secret injection
  * **MCP Duration Strings and Hash Reconciliation** - `tool_sync_interval` accepts Go duration strings; hash-based reconciliation prevents unnecessary MCP client restarts on config reload
  * **Auto-fill Incoming Model for Fallbacks** - Routing rule fallback entries can omit the model; the incoming request model is substituted automatically at runtime
  * **Namespace Tool Type** - Namespace tool container type in Responses API; non-OpenAI providers receive automatically flattened tool lists
  * **Cache Creation Pricing** - Cache creation details for Claude models with 5-minute and 1-hour TTL pricing tiers
  * **Governance Config Sync** - Model configs and provider governance bindings now sync from `config.json` to DB at startup
  * **`business_units` and Team Fields** - Added `business_units`, `team_id`, `calendar_aligned`, and `virtual_key_count` to governance schema and Helm

  ## 🐞 Fixed

  * **WebSocket /responses Reliability** - Fixed upstream handshake diagnostics, proper error capture, and WebSocket connection lifecycle in the native `/responses` path
  * **Raw Request Passthrough Removed** - Removed `SendBackRawRequest` from all provider passthrough flows; passthrough streaming now sets proper SSE headers
  * **WebSocket Nil Checks** - Improved `sendMessageSafely` nil guards, panic recovery, and client cleanup
  * **Routing Rule Query Normalization** - Normalized `query` field to valid `RuleGroupType` and tightened schema validation
  * **PydanticAI Null Text Fields** - Normalized null text content in PydanticAI stream response chunks
  * **Budget and Team Co-creation** - Fixed creation of budgets and teams in the same request
  * **Provider Reload** - Fixed keyless provider status updates during config reload; provider runtime now reloads correctly after key creation
  * **OTel Metrics** - Fixed OpenTelemetry metrics pipeline not working (thanks [@tcx4c70](https://github.com/tcx4c70)!)
  * **OTel Export** - Fixed OTEL exporting to correctly show input and output messages
  * **Logging Request Type** - Resolved request type from pending data before streaming to prevent missing `Object` field in error logs
  * **Multipart File Uploads** - Write multipart metadata before file content to fix upload ordering
  * **Env Var Redacted Check** - Added missing redacted check for env var values
</Update>

<Update label="Core" description="1.5.5">
  * feat: Standardizes tool stripping and anthropic integration handling against anthropic, vertex and azure
  * fix: drops empty thinking block for anthropic provider for claude code
  * feat: add support for cache creation details for claude models
  * feat: adds missing anthropic params / handling in Vertex
  * feat: add namespace tool type support in Responses API
  * feat: support `env.*` references for proxy and TLS config fields
  * feat: add Go duration string support for `tool_sync_interval` and hash-based MCP client config reconciliation
  * fix: add redacted check for env var values
  * fix: write multipart metadata before file uploads
  * fix: removes redundant IsDefined func
  * fix: native WebSocket /responses stream reliability improvements
</Update>

<Update label="Framework" description="1.3.5">
  * fix: Fixes OTEL exporting in `framework/tracing/llmspan.go` to show input and output messages correctly
  * feat: add support for cache creation cost above 1 hour
  * feat: add `provider_key_name` alias for routing targets and pricing overrides, resolved to `key_id` at config load
  * feat: add Go duration string support for MCP `tool_sync_interval` and hash-based client config reconciliation
  * feat: support `env.*` references for proxy and TLS config fields
  * feat: sync model config and provider governance bindings from config file to DB
  * chore: extend migration tests, fix column ignores, add routing rule search
</Update>

<Update label="compat" description="0.1.4">
  * feat: drops tools web\_search and web\_search\_preview if model does not support it
  * feat: flattens tool type namespace if provider does not support it
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="governance" description="1.5.5">
  * feat: automatically add incoming model to empty fallbacks in routing rules
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="jsonparser" description="1.5.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="logging" description="1.5.5">
  * fix: resolve request type from pending data before streaming chunk processing
  * chore: adds debug logging in logging plugin lifecycle
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="maxim" description="1.6.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="mocker" description="1.5.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="otel" description="1.2.5">
  * fix: OTel metrics not working (thanks [@tcx4c70](https://github.com/tcx4c70)!)
  * fix: OTel conversion from built-in trace
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="prompts" description="1.0.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="semanticcache" description="1.5.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>

<Update label="telemetry" description="1.5.5">
  * chore: upgraded core to v1.5.5 and framework to v1.3.5
</Update>