Meridian.Hosting
1.0.0
dotnet add package Meridian.Hosting --version 1.0.0
NuGet\Install-Package Meridian.Hosting -Version 1.0.0
<PackageReference Include="Meridian.Hosting" Version="1.0.0" />
<PackageVersion Include="Meridian.Hosting" Version="1.0.0" />
<PackageReference Include="Meridian.Hosting" />
paket add Meridian.Hosting --version 1.0.0
#r "nuget: Meridian.Hosting, 1.0.0"
#:package Meridian.Hosting@1.0.0
#addin nuget:?package=Meridian.Hosting&version=1.0.0
#tool nuget:?package=Meridian.Hosting&version=1.0.0
Meridian.Hosting
Independent open-source project; not affiliated with or endorsed by Microsoft.
The services.AddMeridian(...) front door — one fluent registration call that wires up
multi-provider models, embeddings/retrieval, agents/tools/memory/workflows, and observability.
Referencing this package alone pulls in the entire Meridian stack
(Meridian.Context,
Meridian.Models,
Meridian.Knowledge,
Meridian.Agents,
Meridian.Observability) — this is the
package most applications should install.
Meridian is an opinionated .NET 10 AI framework built on the Microsoft Agent Framework (MAF). It's a modern .NET framework first — .NET applications reference it directly, no COM involved — and the same engine is also reachable from COM-based applications (VB6, Visual FoxPro, Delphi, VBA) through a COM Callable Wrapper, so a legacy codebase doesn't have to stay frozen out of what's possible with AI today. See the repository for the full story, architecture diagrams, and the COM interop guide.
Install
dotnet add package Meridian.Hosting
Quick start
using Meridian.Hosting;
services.AddMeridian(options => {
options.AddModel(model => {
model.DefaultProvider = ModelProvider.AzureOpenAI;
model.AzureOpenAI.ResourceName = "my-resource";
model.AzureOpenAI.ChatDeploymentName = "gpt-4o";
model.AzureOpenAI.EmbeddingDeploymentName = "text-embedding-3-small";
model.FallbackProviders.Add(ModelProvider.OpenAI);
model.Pipeline.EnableOpenTelemetry = true;
model.Pipeline.EnableRetry = true;
});
options.AddEmbeddings();
options.AddSemanticStore();
options.AddHybridRetrieval(); // vector + keyword fusion; use AddRetrieval() for pure semantic search
options.AddRagContextBuilder(); // batteries-included retrieval -> numbered context -> prompt
options.AddMemory();
options.AddImages(); // generation, editing, vision analysis (AzureOpenAI / OpenAI)
options.AddTool("getWeather", (string city) => $"It's sunny in {city}.");
options.AddAgent("assistant", agent => {
agent.Instructions = "You are a concise, helpful assistant.";
agent.ToolNames.Add("getWeather");
agent.EnableSearchTool = true; // lets the agent search the knowledge base mid-conversation
agent.EnableTracing = true;
});
});
var agent = serviceProvider.GetRequiredService<IMeridianAgentRegistry>().Get("assistant");
var session = await agent.CreateSessionAsync();
var response = await agent.RunAsync("What's the weather in Seattle?", session);
Console.WriteLine(response.Text);
Feature surface
| Capability | How |
|---|---|
| Stateless chat | IChatClient (registered by AddModel) |
| Multi-turn chat with history | IMeridianAgentRegistry → AIAgent + session (MAF) |
| Embeddings | IEmbeddingService (AddEmbeddings) |
| Semantic (vector) search | AddRetrieval() → IRetriever |
| Hybrid search (vector + keyword, weighted fusion) | AddHybridRetrieval() → IRetriever |
| Batteries-included RAG (retrieval + context + prompt template + sources) | AddRagContextBuilder() → RagContextBuilder |
| Long-term agent memory (remember/recall/forget) | AddMemory() → IMeridianMemory |
| Image generation / editing / vision analysis | AddImages() → IMeridianImageService |
| Agent tools (custom + built-in knowledge search) | AddTool(...), MeridianAgentOptions.EnableSearchTool |
| Sequential workflows (agent + transform steps) | AddWorkflow(...) → MeridianWorkflowBuilder |
| Tracing / metrics / evaluation | Meridian.Observability (MeridianTelemetry, MeridianMetrics, IEvaluator) |
Switch model providers — or add fallback providers tried in order — by changing ModelOptions
inside AddModel(...); no other code in your application changes. Supported providers: Azure
OpenAI, OpenAI, Anthropic (Claude), Google Gemini, Perplexity, and Azure AI Inference (Foundry).
Reaching it from COM (VB6, Visual FoxPro, Delphi, VBA)
The same engine configured via AddMeridian(...) is exposed through Meridian.Interop.Com's five
COM-creatable classes — Meridian.ChatService, Meridian.AgentService,
Meridian.EmbeddingService, Meridian.MemoryService, Meridian.ImageService — callable from any
COM host in a few lines:
Dim chat As Object
Set chat = CreateObject("Meridian.ChatService")
Debug.Print chat.Chat("What's 2 + 2?")
Meridian.Interop.Com isn't published to NuGet (it's a RID-specific dotnet publish artifact,
not a library you reference) — see
COM-INTEROP.md in the
repository for the full registration and calling guide.
License
MIT © Angel Hernandez
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net10.0
- Meridian.Agents (>= 1.0.0)
- Meridian.Context (>= 1.0.0)
- Meridian.Knowledge (>= 1.0.0)
- Meridian.Models (>= 1.0.0)
- Meridian.Observability (>= 1.0.0)
- Microsoft.Extensions.DependencyInjection.Abstractions (>= 10.0.4)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 1.0.0 | 100 | 10/13/2026 |