Quickstart

This quickstart takes you from a new account to a working API call, then to a connected coding agent. You need a terminal and, for the SDK examples, a recent Python 3 or Node.js 18+.

1. Create an account#

Sign up at tokens.bd/register with email or Google. Verify your email address: you can't create API keys until it is verified.

2. Add credit: a plan or a top-up#

Keys only work when there is something to pay for usage with. Pick one (or both) in /dashboard/billing:

• Subscribe to a plan for a weekly or monthly credit allowance. Compare them on pricing.
• Top up your wallet for pay-as-you-go use. The minimum top-up is $5 or ৳500.

Prices are shown in USD, BDT. Available payment methods: Invoice via billing portal. If you want to pay in taka, read paying in BDT first.

3. Create an API key#

Open /dashboard/keys and create a key. Give it a name you will recognise later ("laptop", "ci", "staging-bot"). You can also set a monthly spend cap and limit the key to certain models; both are optional, and both are fixed once the key exists. API keys explains when to use them.

The secret starts with tok_live_ and is shown once. Copy it now and put it in an environment variable:

export TOKENS_API_KEY="tok_live_your_key"

$env:TOKENS_API_KEY = "tok_live_your_key"

4. Make your first request#

The OpenAI-compatible base URL is https://tokens.bd/v1. The examples use deepseek/deepseek-v4.1-flash; swap in any model id from /models that your plan or wallet covers.

curl https://tokens.bd/v1/chat/completions \
  -H "Authorization: Bearer $TOKENS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4.1-flash",
    "messages": [{"role": "user", "content": "Say hello in one short sentence."}]
  }'

# pip install openai
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://tokens.bd/v1",
    api_key=os.environ["TOKENS_API_KEY"],
)

response = client.chat.completions.create(
    model="deepseek/deepseek-v4.1-flash",
    messages=[{"role": "user", "content": "Say hello in one short sentence."}],
)
print(response.choices[0].message.content)

// npm install openai   (save as hello.mjs, run: node hello.mjs)
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://tokens.bd/v1",
  apiKey: process.env.TOKENS_API_KEY,
});

const response = await client.chat.completions.create({
  model: "deepseek/deepseek-v4.1-flash",
  messages: [{ role: "user", content: "Say hello in one short sentence." }],
});
console.log(response.choices[0].message.content);

A successful response is a normal OpenAI chat completion. If you get an error instead, the error.code field tells you why:

The full list is in errors.

List the models your key can use#

curl https://tokens.bd/v1/models -H "Authorization: Bearer $TOKENS_API_KEY"

The id of each entry in data is exactly what you put in model.

5. Connect your coding agent#

The quickest route for OpenCode, Claude Code, Codex CLI and Crush is the Tokens CLI. It signs you in through the browser, lets you pick a default model, and edits the agents' config files with a backup of each:

curl -fsSL https://tokens.bd/cli/tokens.mjs -o tokens.mjs && node tokens.mjs setup --base-url https://tokens.bd

iwr https://tokens.bd/cli/tokens.mjs -OutFile tokens.mjs; node tokens.mjs setup --base-url https://tokens.bd

See Tokens CLI for what it changes. If you prefer to configure things by hand, or use another tool, /dashboard/connect has copy-and-paste settings for each agent, and the guides cover them in detail:

• Claude Code (Anthropic-compatible, base URL https://tokens.bd)
• Codex CLI, OpenCode, Crush
• Cursor, Cline, Roo Code, Kilo Code, Continue, Zed, Aider

Test your key from this page#

Paste your key below to send a short test request through the gateway. It costs a few tokens.

Next steps#

• Choosing a model: which model for which job.
• Usage, limits and alerts: watch spend and set up alerts before you hand a key to an agent.
• Streaming and tool calling for the API details agents rely on.