BlazorMemory v0.8.0 dropped this week, and the headline feature is exactly what the community has been demanding since day one: a proper Semantic Kernel adapter. The library now implements Microsoft's IMemoryStore interface directly, which means .NET developers can drop BlazorMemory into existing SK-based pipelines without fighting custom integration code.

Semantic Kernel Adapter Ships at Last

The long-awaited compatibility layer is available as the BlazorMemory.SemanticKernel NuGet package. Configuration follows familiar DI patterns: developers wire up .AddBlazorMemory() chained with .UseInMemoryStorage(), and the IMemoryStore implementation handles the rest. This is a clean win for teams building agentic applications on .NET who want persistent conversation context without vendor lock-in.

Ollama Embeddings Enter the Picture

Beyond SK integration, v0.8.0 adds native support for Ollama embeddings—critical for developers running models locally rather than hitting OpenAI or Azure endpoints. The ability to generate and store vector representations through an Ollama backend brings BlazorMemory fully into the local-first AI development workflow that many .NET shops are now targeting.

Memory Decay Joins the Feature List

The release also includes a memory decay mechanism, addressing one of the persistent challenges in long-running conversational applications: how to manage context without letting stale information pollute future interactions. While specific configuration options weren't detailed in the initial announcement, this suggests BlazorMemory is evolving beyond simple retrieval toward more sophisticated memory management patterns.

Key Takeaways

  • Semantic Kernel IMemoryStore implementation opens cross-compatibility with Microsoft's agent framework
  • Ollama embeddings support enables fully local deployment without cloud API dependencies
  • Memory decay provides a path to managing context quality over extended conversations
  • Package BlazorMemory.SemanticKernel available now on NuGet

The Bottom Line

This release signals BlazorMemory's maturation from an interesting experiment into a production-grade component for .NET AI applications. The SK adapter alone makes it worth revisiting if you've been managing conversation memory manually—because let's be honest, you probably weren't doing it as well as this library does.