Quickstart

Go from zero to your first token-hub API call in under five minutes. Sign up, create a key, send a cURL request, and connect OpenAI-compatible or Anthropic Messages-compatible clients.

token-hub is an LLM API gateway with OpenAI Chat Completions, Anthropic Messages, and Responses-compatible routes behind one TokenHub key. This guide gets you from sign-up to a live response using an enabled model route.

Open tokenhub.sandboxclaw.com/register and create an account with email. No phone number is required.

Verify your email if the console asks you to. That unlocks the dashboard.

Step 2 - Top up and create a key

From the dashboard, go to /topup. Public checkout is currently USD-only through Stripe, with topups starting at $5. Your card issuer or bank may handle any local currency conversion.

Then visit /keys and click New Key. Copy the full TokenHub key shown by the console. We hash the key server-side and cannot show it again; if you lose it, revoke and reissue.

Step 3 - Send your first OpenAI-compatible request

The OpenAI-compatible base URL is https://llm.sandboxclaw.com/v1.

curl https://llm.sandboxclaw.com/v1/chat/completions \
  -H "Authorization: Bearer $TOKENHUB_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshot-v1-8k",
    "messages": [{"role": "user", "content": "Reply exactly ok"}],
    "max_tokens": 8
  }'

You get back an OpenAI-shaped response:

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "model": "moonshot-v1-8k",
  "choices": [
    { "index": 0, "message": { "role": "assistant", "content": "ok" }, "finish_reason": "stop" }
  ],
  "usage": { "prompt_tokens": 12, "completion_tokens": 1, "total_tokens": 13 }
}

The usage block drives billing. Token counts come from the upstream channel and are passed through.

Step 4 - Use the OpenAI Python SDK

Override base_url and pass the TokenHub key:

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://llm.sandboxclaw.com/v1",
    api_key=os.environ["TOKENHUB_KEY"],
)

resp = client.chat.completions.create(
    model="moonshot-v1-8k",
    messages=[{"role": "user", "content": "Reply exactly ok"}],
    max_tokens=8,
)
print(resp.choices[0].message.content)

For streaming, set "stream": true. The response is a text/event-stream of OpenAI-compatible delta chunks terminated by data: [DONE].

Step 5 - Connect Claude Code-style clients

For Anthropic Messages-compatible clients, use the Anthropic-style base URL without /v1:

export ANTHROPIC_BASE_URL=https://llm.sandboxclaw.com
export ANTHROPIC_API_KEY=$TOKENHUB_KEY

claude --bare -p --model moonshot-v1-8k "Reply exactly ok"

Claude Code reaches /v1/messages with this configuration.

Step 6 - Know the current CLI status

Claude Code: set ANTHROPIC_BASE_URL=https://llm.sandboxclaw.com and ANTHROPIC_API_KEY=$TOKENHUB_KEY.
OpenCode: configure an OpenAI-compatible provider with baseURL: "https://llm.sandboxclaw.com/v1" and model moonshot-v1-8k.
DeepSeek-TUI: set DEEPSEEK_API_KEY=$TOKENHUB_KEY, DEEPSEEK_BASE_URL=https://llm.sandboxclaw.com/v1, and DEEPSEEK_MODEL=moonshot-v1-8k.
Codex: configure TokenHub as a Responses provider. Model availability follows the API key’s enabled model list.