LangChain talks to providers through its chat models. Point those models at the Gateway the same way you would point the native SDKs. LangGraph uses the same ChatOpenAI / ChatAnthropic objects.

What this Gateway accepts

OpenAI: POST /openai/v1/chat/completions and GET /openai/v1/models. Anthropic: POST /anthropic/v1/messages. Gemini Developer API: POST /google/{v1beta|v1}/models/{model}:generateContent and :streamGenerateContent. Embeddings, the OpenAI Responses API, Assistants, Azure OpenAI paths, Vertex AI paths, and Anthropic Batches are not this Gateway. Those requests 404.

Headers

Header	Value
X-Traccia-Api-Key	Workspace key from Settings → API Keys
X-Traccia-Agent-Id	Same Agent ID you would pass to init

Keep the provider API key on the client. The Gateway forwards it and does not store it. Missing Traccia key returns 401. Missing agent ID returns 400.

ChatOpenAI

Official knobs: base_url (alias openai_api_base) and default_headers. Resolution is explicit base_url, then OPENAI_API_BASE, then OPENAI_BASE_URL. The value must be https://gateway.traccia.ai/openai/v1.

Force Chat Completions

Newer langchain-openai can post to /v1/responses from the model name, output_version, or Responses-only features. This Gateway only implements POST /openai/v1/chat/completions. Set use_responses_api=False in Python and useResponsesApi: false in JavaScript. Do not use Responses-only tools or kwargs.

python

import os
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="gpt-4o-mini",
    api_key=os.environ["OPENAI_API_KEY"],
    base_url="https://gateway.traccia.ai/openai/v1",
    default_headers={
        "X-Traccia-Api-Key": os.environ["TRACCIA_API_KEY"],
        "X-Traccia-Agent-Id": "billing-bot",
    },
    use_responses_api=False,
)

response = llm.invoke("hello")

JavaScript puts OpenAI SDK options under configuration (baseURL and defaultHeaders). Some LangChain JS versions still route to /v1/responses when tools are present even if useResponsesApi is false. If you see that 404, drop Responses-only tools or call Chat Completions through the OpenAI SDK page.

ChatAnthropic

Official knobs: base_url (alias of anthropic_api_url) and default_headers. The Anthropic SDK default is https://api.anthropic.com, so the Gateway value is https://gateway.traccia.ai/anthropic with no /v1.

python

import os
from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic(
    model="claude-sonnet-4-5",
    api_key=os.environ["ANTHROPIC_API_KEY"],
    base_url="https://gateway.traccia.ai/anthropic",
    default_headers={
        "X-Traccia-Api-Key": os.environ["TRACCIA_API_KEY"],
        "X-Traccia-Agent-Id": "billing-bot",
    },
)

response = llm.invoke("hello")

LangSmith Gateway Is Not Traccia

LangChain can fall back to LANGSMITH_GATEWAY. That is LangSmith, not this product. Set base_url explicitly to the Traccia host.

Not Documented Here

ChatGoogleGenerativeAI is not verified against this Gateway. Use the Gemini SDK page for Gemini traffic.

Gateway: LangChain

Headers

ChatOpenAI

ChatAnthropic

Not Documented Here

Next Steps

Gateway Overview

Policies