Fairmeter
All docs

Reference

SDK recipes

Runnable templates for each supported SDK. Drop the gateway base URL into an existing project, set one env var, run the snippet. The response carries the gateway's X-Request-Id header so you can correlate any result back to the routing decision in your dashboard.

OpenAI · Python

Drop-in OpenAI(base_url=...) swap with X-Request-Id correlation.

python
from openai import OpenAI

client = OpenAI(
    base_url="https://gateway.fairmeter.in/v1",
    api_key="gw_live_xxxxxxxx",
)

resp = client.chat.completions.create(
    model="auto",
    messages=[{"role": "user", "content": "Summarise this contract."}],
)
print(resp.id, resp.model)

OpenAI · Node

TypeScript client config with the gateway base URL plus retry config.

typescript
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://gateway.fairmeter.in/v1",
  apiKey: process.env.FAIRMETER_API_KEY,
  maxRetries: 2,
});

const resp = await client.chat.completions.create({
  model: "auto",
  messages: [{ role: "user", content: "Summarise this contract." }],
});

Anthropic · Python

Anthropic SDK with base_url swap; pinned claude-haiku-4-5.

python
from anthropic import Anthropic

client = Anthropic(
    base_url="https://gateway.fairmeter.in",
    api_key="gw_live_xxxxxxxx",
)

message = client.messages.create(
    model="claude-haiku-4-5",
    max_tokens=256,
    messages=[{"role": "user", "content": "Say hi."}],
)

Anthropic · Node

@anthropic-ai/sdk config with the gateway origin and Haiku pin.

typescript
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://gateway.fairmeter.in",
  apiKey: process.env.FAIRMETER_API_KEY,
});

const message = await client.messages.create({
  model: "claude-haiku-4-5",
  max_tokens: 256,
  messages: [{ role: "user", content: "Say hi." }],
});

curl

Two-flag swap: Authorization: Bearer plus the gateway origin.

bash
curl https://gateway.fairmeter.in/v1/chat/completions \
  -H "Authorization: Bearer gw_live_xxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{"model":"auto","messages":[{"role":"user","content":"hi"}]}'

LangChain · Python

ChatOpenAI and ChatAnthropic configured against the gateway.

python
from langchain_openai import ChatOpenAI
from langchain_anthropic import ChatAnthropic

llm = ChatOpenAI(
    base_url="https://gateway.fairmeter.in/v1",
    api_key="gw_live_xxxxxxxx",
    model="auto",
)

claude = ChatAnthropic(
    anthropic_api_url="https://gateway.fairmeter.in",
    anthropic_api_key="gw_live_xxxxxxxx",
    model="claude-haiku-4-5",
)

Cursor

OpenAI Base URL Override field in Cursor’s Models settings panel.

Open Cursor Settings → Models → OpenAI API Key, set the key to your gw_live_ token, and set “OpenAI Base URL” (under Override) to https://gateway.fairmeter.in/v1. Cursor’s own model picker keeps working; every request now routes through the gateway, and a concrete model id sent by Cursor triggers the family lock described in Routing.

Codex CLI

Custom model_provider in ~/.codex/config.toml pointed at the gateway.

toml
# ~/.codex/config.toml
[model_providers.fairmeter]
name = "Fairmeter"
base_url = "https://gateway.fairmeter.in/v1"
env_key = "FAIRMETER_API_KEY"

model_provider = "fairmeter"
model = "auto"

Claude Code

ANTHROPIC_BASE_URL plus ANTHROPIC_AUTH_TOKEN; family-lock keeps Anthropic.

bash
export ANTHROPIC_BASE_URL="https://gateway.fairmeter.in"
export ANTHROPIC_AUTH_TOKEN="gw_live_xxxxxxxx"

claude

Claude Code always sends a concrete Anthropic model id, so the family lock in Routing keeps every step of the session on the Anthropic ladder while still routing each step to the tier the classifier picks.

Every snippet above uses the same gateway origin and a placeholder key. Swap in your own gw_live_ key from /keys and the response carries an X-Request-Id header you can look up at /requests/<id>.
PreviousWebhook capacityNext API reference