Quickstart: LangChainGo with Anthropic

Get started by running your first program with LangChainGo and Anthropic. Anthropic's Claude models are known for strong reasoning, long context windows, and extended ("thinking") modes.

Prerequisites

1. Anthropic API Key: Sign up on the Anthropic Console and retrieve your API key.
2. Go: Download and install Go.

Setup

Before interacting with the Anthropic API, you need to set up your API key as an environment variable.

Linux/macOS (bash/zsh)

export ANTHROPIC_API_KEY="your_anthropic_api_key_here"

Windows (Command Prompt)

set ANTHROPIC_API_KEY=your_anthropic_api_key_here

Windows (PowerShell)

$env:ANTHROPIC_API_KEY="your_anthropic_api_key_here"

For permanent setup, add the environment variable to your shell's profile file (~/.bashrc, ~/.zshrc, etc.) or system environment variables on Windows.

Steps

1. Set up your Anthropic API Key: Follow the setup instructions above to configure your API key.

2. Run the example: Execute the following command:

go run github.com/vxcontrol/langchaingo/examples/anthropic-completion-example@main-vxcontrol

The example streams the model's response chunk by chunk, so you will see a poem about Go-powered AI systems printed as it is generated.

Congratulations! You have successfully built and executed your first LangChainGo LLM-backed program using Anthropic's cloud-based inference.

Selecting a model

The client picks a sensible default model, but you can choose one explicitly with anthropic.WithModel:

llm, err := anthropic.New(anthropic.WithModel("claude-sonnet-4-5-20250929"))

Extended thinking

Claude's extended-thinking models can spend extra tokens reasoning before answering. Enable it per call with the provider-neutral reasoning options — llms.WithReasoning for a token budget or effort, or llms.WithAdaptiveReasoning on the newest generations. The adapter resolves the correct wire shape (budget vs. adaptive) from the model, so a request is never sent in a form the model rejects:

completion, err := llms.GenerateFromSinglePrompt(ctx, llm, prompt,
    llms.WithReasoning(llms.ReasoningMedium, 0),
)

See Configure LLM Providers for the full set of reasoning and structured-output options.

Here is the entire program (from anthropic-completion-example):

package main

import (
	"context"
	"fmt"
	"log"

	"github.com/vxcontrol/langchaingo/llms"
	"github.com/vxcontrol/langchaingo/llms/anthropic"
	"github.com/vxcontrol/langchaingo/llms/streaming"
)

func main() {
	llm, err := anthropic.New(
		anthropic.WithModel("claude-3-5-sonnet-20240620"),
	)
	if err != nil {
		log.Fatal(err)
	}
	ctx := context.Background()
	completion, err := llms.GenerateFromSinglePrompt(ctx, llm, "Hi claude, write a poem about golang powered AI systems",
		llms.WithTemperature(0.8),
		llms.WithStreamingFunc(func(_ context.Context, chunk streaming.Chunk) error {
			fmt.Println(chunk.String())
			return nil
		}),
	)
	if err != nil {
		log.Fatal(err)
	}

	_ = completion
}