> ## 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.

# v1.5.15

> v1.5.15 changelog - 2026-06-16

<Tabs>
  <Tab title="NPX">
    ```bash theme={null}
    npx -y @maximhq/bifrost --transport-version v1.5.15
    ```
  </Tab>

  <Tab title="Docker">
    ```bash theme={null}
    docker pull maximhq/bifrost:v1.5.15
    docker run -p 8080:8080 maximhq/bifrost:v1.5.15
    ```
  </Tab>
</Tabs>

<Update label="Bifrost(HTTP)" description="1.5.15">
  ## ✨ Features

  * **pprof Profiling Server** - Optional runtime profiling server gated by `BIFROST_PPROF_PORT`, with env-tunable block/mutex sampling rates and graceful shutdown alongside the main server.
  * **Anthropic Cache Diagnostics** - Surface Anthropic's prompt-cache diagnostics beta (`cache-diagnosis-2026-04-07`): responses now expose the first cache-prefix divergence point, so you can see exactly why a prompt cache missed.

  ## 🐞 Fixed

  * **GenAI Raw Request Passthrough** - Native Vertex/Gemini batch and request bodies now follow the `x-model-provider` header and pass through verbatim only when Gemini or Vertex is explicitly selected, preventing a raw body from reaching a mismatched provider.
  * **Tool Call Metadata Preservation** - `extra_content` on assistant tool calls (e.g. Gemini `thought_signature`) is now preserved across both streaming and non-streaming responses (thanks [@nghodkicisco](https://github.com/nghodkicisco)!).
</Update>

<Update label="Core" description="1.5.21">
  * fix: preserve extra\_content on ChatAssistantMessageToolCall so provider-specific tool-call metadata (e.g. Gemini thought\_signature) round-trips losslessly through the gateway (thanks [@nghodkicisco](https://github.com/nghodkicisco)!)
  * fix: surface Anthropic prompt-cache diagnostics (cache-diagnosis-2026-04-07 beta) via new CacheDiagnostics/CacheMissReason schema, reporting the first cache-prefix divergence point on chat and responses APIs
</Update>

<Update label="Framework" description="1.3.21">
  * fix: carry tool-call extra\_content through the streaming accumulator - deep-copied per chunk and accumulated across deltas (thanks [@nghodkicisco](https://github.com/nghodkicisco)!)
  * chore: upgraded core to v1.5.21
</Update>

<Update label="compat" description="0.1.20">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="governance" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="jsonparser" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="logging" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="maxim" description="1.6.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="mocker" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="modelcatalogresolver" description="1.0.2">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="otel" description="1.2.22">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="prompts" description="1.0.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="semanticcache" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>

<Update label="telemetry" description="1.5.21">
  * chore: upgraded core to v1.5.21 and framework to v1.3.21
</Update>