Quickstart

Supported target frameworks: .NET Standard 2.0, .NET 6, or .NET 8.
An LLM endpoint. Built-in support for Ollama, OpenAI, Anthropic, Gemini, Azure OpenAI, AWS Bedrock, and any OpenAI-compatible endpoint (vLLM, TEI, LM Studio, DeepSeek, Groq). You can also plug in a custom provider by implementing LlmClientBase.

LogicGrid is provider-agnostic. Swap one line to switch providers — see Providers.

1. Install

dotnet new console -n MyAgentApp
cd MyAgentApp
dotnet add package LogicGrid.Core

LogicGrid.Core is the only package you need to get started. It covers agents, admins, all orchestration patterns, all LLM providers, events, logging, and tracing. Add more packages when you need them:

Package	Purpose
`LogicGrid.Core`	Agents, admins, providers, orchestration
`LogicGrid.Tools`	Built-in tools (calculator, HTTP, web search)
`LogicGrid.Mcp`	Model Context Protocol client
`LogicGrid.Memory`	Vector stores, embeddings, agent memory
`LogicGrid.Rag`	RAG pipeline, hybrid search, document loaders

2. Your first agent

Replace the contents of Program.cs:

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);

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

Paris.

That's the whole flow: pick an LLM, build an agent with a name and a system prompt, call RunAsync.

3. Add observability

LogicGrid ships logging and tracing as fluent extensions on AgentContext — designed for simplicity:

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

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);

var ctx = new AgentContext()
    .WithLogging()
    .WithTracing(out var trace);

var result = await agent.RunAsync("Capital of France?", ctx);

Console.WriteLine($"\nResult : {result}");
Console.WriteLine($"Tokens : {trace.TotalTokenUsage.TotalTokens}");
Console.WriteLine($"Cost   : ${trace.TotalCost.TotalCostUsd:0.0000}");
Console.WriteLine($"Time   : {trace.Duration.TotalMilliseconds:0}ms");

09:14:02.118 [INF] [a3f2c891] [Helper] started | input: Capital of France?
09:14:02.120 [DBG] [a3f2c891] [Helper] LLM call started | model: llama3.2
09:14:03.398 [DBG] [a3f2c891] [Helper] LLM call completed | 1278ms | 18 tokens
09:14:03.402 [INF] [a3f2c891] [Helper] completed | output: Paris. | 1284ms | 18 tokens

Result : Paris.
Tokens : 18
Cost   : $0.0000
Time   : 1284ms

WithLogging provides live console feedback, while WithTracing captures a comprehensive telemetry snapshot for analysis — Spans, token usage, cost, retries, and tool calls.

Logger output can be tuned via LogicGridLoggerOptions (verbosity, sinks, what fields to show) — see Logger options.

4. Orchestrate with an admin

An admin coordinates several agents. LogicGrid provides several implementations — including SequentialAdmin, GroupChatAdmin, GraphAdmin, ParallelAdmin, MapReduceAdmin, and ReflexionAdmin. each exposing a standard RunAsync method. The example below uses SequentialAdmin:

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

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

IAgent summary = new Agent<string>(
    name:         "Summary",
    description:  "Summarises text in 2-3 sentences.",
    systemPrompt: "You are a concise summariser. Answer in 2-3 sentences.",
    llm:          llm);

IAgent review = new Agent<string>(
    name:         "Review",
    description:  "Scores a summary 1-10.",
    systemPrompt: "Score the summary from 1 to 10 and explain in one sentence.",
    llm:          llm);

var admin = new SequentialAdmin<string, string>(
    name:      "Editorial",
    llmClient: llm,
    agents:    new[] { summary, review });

var ctx = new AgentContext().WithLogging();
var output = await admin.RunAsync(
    input: "The James Webb Space Telescope launched in December 2021. " +
           "It observes infrared light from the earliest galaxies and " +
           "studies exoplanet atmospheres.");

Console.WriteLine($"\n{output}");

14:11.080 [INF] [b7d4e210] Run started — admin: Editorial | task: The James Webb Space Telescope launched in December 2021. …
14:11.220 [INF] [b7d4e210] [Summary] started | input: The James Webb Space Telescope launched in December 2021. …
14:14.402 [INF] [b7d4e210] [Summary] completed | output: The James Webb Space Telescope, launched in 2021, observes infrared light … | 3182ms | 188 tokens
14:14.430 [INF] [b7d4e210] [Review] started | input: The James Webb Space Telescope, launched in 2021, observes infrared light …
14:16.871 [INF] [b7d4e210] [Review] completed | output: 9/10. Concise and accurate; could mention exoplanet atmospheres explicitly. | 2441ms | 102 tokens
14:16.872 [INF] [b7d4e210] Run completed — 2 agents, 2 LLM calls | 5792ms

9/10. Concise and accurate; could mention exoplanet atmospheres explicitly.

Where next

Orchestration — Sequential, Group chat, Graph, Parallel, Map-reduce, Reflexion. Each page has a diagram and a runnable example.
Tools — give agents access to web search, HTTP, math, JSON extraction, custom tools: Tools.
Memory + RAG — embed documents, retrieve them, ask questions: RAG pipeline.
Customise — when you need to override prompt rendering, output validation, or retry behaviour: Advanced.
A different provider — Providers.

1. Install​

2. Your first agent​

3. Add observability​

4. Orchestrate with an admin​

Where next​

1. Install

2. Your first agent

3. Add observability

4. Orchestrate with an admin

Where next