On this page
PrerequisitesSetupBasic Chat CompletionSwitching ModelsStreamingConclusionRelated
Tutorials

Access Any Model Using LangChain

Reynaldi Chernando
On this page
PrerequisitesSetupBasic Chat CompletionSwitching ModelsStreamingConclusionRelated

In this tutorial, you'll learn how to access any AI model using LangChain through Puter's OpenAI-compatible endpoint. GPT, Claude, Gemini, Grok, DeepSeek, Llama, and more, all through a single endpoint. No separate API keys needed, just your Puter auth token.

Prerequisites

Puter copy auth token
  • Python 3.9+ installed on your machine
  • uv installed on your machine

Setup

Create a new project and install langchain-openai:

uv init puter-langchain
cd puter-langchain
uv add langchain-openai

Then configure the client with Puter's base URL and your auth token:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="gpt-5-nano",
)

Replace YOUR_PUTER_AUTH_TOKEN with the auth token you copied from your Puter dashboard. That's all you need. No separate API keys required for any model.

Basic Chat Completion

Here's a simple chat completion using GPT:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="gpt-5-nano",
)

response = llm.invoke("What is the capital of France?")
print(response.text)

Sample output:

The capital of France is Paris.

The code is identical to what you'd write for OpenAI directly. The only difference is the base URL and auth token. Run it with:

uv run main.py

Switching Models

This is where it gets interesting. Same code, same setup. Just change the model parameter to use any supported model:

from langchain_openai import ChatOpenAI

# Use Claude
claude = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="claude-sonnet-4-5",
)
print("Claude:", claude.invoke("What is the capital of France?").text)

# Use Gemini
gemini = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="gemini-2.5-flash-lite",
)
print("Gemini:", gemini.invoke("What is the capital of France?").text)

# Use Grok
grok = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="grok-4-1-fast",
)
print("Grok:", grok.invoke("What is the capital of France?").text)

One endpoint, any model. You don't need separate SDKs, separate API keys, or separate billing accounts. Switch between providers by changing a single string.

Streaming

For longer responses, streaming gives you results in real-time as they're generated:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
    model="claude-sonnet-4-5",
)

for chunk in llm.stream("Write a short story about a robot learning to paint."):
    print(chunk.text, end="", flush=True)

Use stream instead of invoke and iterate over the chunks as they arrive. Each chunk contains a piece of the response that you can display immediately. This works with any model.

Conclusion

That's it. You now have LangChain connected to any AI model through Puter, GPT, Claude, Gemini, Grok, and more through a single endpoint. No need to juggle multiple API keys or rewrite your code when you want to try a different model. Just swap the model string.

Free, Serverless AI and Cloud

Start creating powerful web applications with Puter.js in seconds!

Get Started Now

Read the Docs • Try the Playground