LogicGrid

Multi-agent AI orchestration for .NET

Local LLMs first. Provider-agnostic. Zero vendor lock-in. Build agents with Agent<T>, compose them through admins, observe everything by default.

Get Started Free View on GitHub →

using LogicGrid.Core.Agents;
using LogicGrid.Core.Llm;

var llm = LlmClientBase.Ollama("llama3.2");

IAgent agent = new Agent<string>(
    name: "Helper",
    description: "Answers questions concisely.",
    systemPrompt: "Answer in one short sentence.",
    llm: llm);

Console.WriteLine(
    await agent.RunAsync("Capital of France?", new AgentContext()));

Why LogicGrid

◧

Works with any LLM

Ollama, Claude, OpenAI, Gemini, vLLM, TEI, LM Studio, DeepSeek, Groq. Same code, any provider. Swap models in one line.

◉

See everything

Structured events on every run. .WithLogging().WithTracing(out var trace) gives you token counts, costs, retries, tool calls — no boilerplate, no black boxes.

▥

Pick the pattern

Sequential, group chat, parallel, map-reduce, reflexion, graph. Each is a single class. Compose them, swap them, or customize your own.

⛨

Built for security

Run entirely on local LLMs — no data ever leaves your network. API keys stay in your process and are never logged. No telemetry. The framework is yours to audit.

Real code, real output

Every example below is runnable. Output is verbatim from the corresponding sample.

Code

using LogicGrid.Core.Agents;
using LogicGrid.Core.Llm;

var llm = LlmClientBase.Ollama("llama3.2");

IAgent agent = new Agent<string>(
    name: "Helper",
    description: "A helpful assistant.",
    systemPrompt: "Answer the user concisely.",
    llm: llm);

var ctx = new AgentContext();
Console.WriteLine(await agent.RunAsync("Say hello in three languages.", ctx));

Console output

Hello (English) — Bonjour (French) — Hola (Spanish)