Your first app
Moth DocsTutorials
A small program in Python, TypeScript or curl that asks a model one question, prints the answer and says what the call cost.
Get a key
Make one in Moth Console under Keys, as the Quickstart shows. It starts with
moth_, and the browser you made it in keeps a copy you can copy again.Put it in your shell
Keep the key out of your code by setting it as
MOTH_API_KEY, here for this shell only.Install the SDK
Moth speaks the OpenAI API, so the official OpenAI SDK is all you need. Python needs version 3.8 or later and TypeScript needs Node 18 or later.
Run the program
Copy the file for your language below and run it. The answer prints first, then the cost.
export MOTH_API_KEY=moth_your_key_here pip install openai # for Python npm install openai tsx # for TypeScript
Python
import os from openai import OpenAI client = OpenAI(base_url="https://api.runmoth.si/v1", api_key=os.environ["MOTH_API_KEY"]) raw = client.chat.completions.with_raw_response.create( model="gemini-3.8-flash", messages=[{"role": "user", "content": "In one sentence, why do moths fly to lamps?"}], ) reply = raw.parse() print(reply.choices[0].message.content) print("Cost in dollars", int(raw.headers["moth-charged-micros"]) / 1_000_000)
Run it with python first.py.
TypeScript
import OpenAI from "openai"; const apiKey = process.env.MOTH_API_KEY; if (!apiKey) throw new Error("Set MOTH_API_KEY"); const client = new OpenAI({ baseURL: "https://api.runmoth.si/v1", apiKey }); const { data: reply, response } = await client.chat.completions .create({ model: "gemini-3.8-flash", messages: [{ role: "user", content: "In one sentence, why do moths fly to lamps?" }], }) .withResponse(); console.log(reply.choices[0].message.content); console.log("Cost in dollars", Number(response.headers.get("moth-charged-micros")) / 1_000_000);
Run it with npx tsx first.mts. The .mts ending lets the file use await at the top level.
curl
curl -i https://api.runmoth.si/v1/chat/completions \ -H "Authorization: Bearer $MOTH_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "gemini-3.8-flash", "messages": [{"role": "user", "content": "In one sentence, why do moths fly to lamps?"}]}'
The -i prints the reply's headers above its body, so you can see moth-charged-micros beside the answer.
What came back
The body is the usual chat completion, with the answer under choices and the token counts under usage. The moth-charged-micros header is what the call cost in millionths of a dollar, and Receipts and headers lists the other headers.
If it fails
MOTH_API_KEY in this shell, or make a new keyretry-after, then try againErrors and rate limits lists every error. Next, stream the answer into a web page.