Getting started

From zero to your first AI response in about five minutes — no waitlist, no sales call, no separate account for every model you want to try.

1. Create an account

Go to https://tokenbroker.hopto.org and sign up. You can use:

New accounts start with a wallet. To use paid models you’ll top up with a card or Cash App (processed through Square), or buy a plan.

2. Create an API key

From the dashboard:

  1. Open API Keys.
  2. Click Create key.
  3. Copy the key and store it somewhere safe. You can set a daily spend cap on each key, and you can revoke a key at any time.

3. Make your first request

The API is OpenAI-compatible. Point any OpenAI client at the service with your key:

curl https://tokenbroker.hopto.org/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-oss:20b-cloud",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

Or in Python:

import openai

client = openai.OpenAI(
    base_url="https://tokenbroker.hopto.org/v1",
    api_key="YOUR_API_KEY",
)

resp = client.chat.completions.create(
    model="gpt-oss:20b-cloud",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(resp.choices[0].message.content)

You can also stream responses, and the same key works for embeddings and text completions. Streaming looks like this in Python:

stream = client.chat.completions.create(
    model="gpt-oss:20b-cloud",
    messages=[{"role": "user", "content": "Write a haiku about compute."}],
    stream=True,
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

4. Browse the catalog

The dashboard’s Models page lists every available model with:

See the models catalog for what’s available.

5. Try the playground

No API client handy? The Playground lets you chat with any model directly in the browser, see the calculated cost of each exchange, and copy the exact code for what you just did.

6. Check your usage

The dashboard shows wallet balance, today’s spend, and per-key usage. The API also exposes a usage endpoint so your own tooling can track it.

Troubleshooting your first request

Symptom Likely cause Fix
401 immediately Key typo, or the key was revoked Copy the key again from the dashboard; keys are hashed at rest so they can’t be “looked up,” only re-shown at creation
402 Empty wallet, spend cap hit, or a minimum-balance requirement on some models Top up, or raise the key’s daily spend cap
404 model_not_found Model name typo, or the model was retired upstream Check the exact spelling on the Models page — provider names change occasionally, see Tested & verified
429 Per-key rate limit (RPM/TPM) or daily spend cap Slow down, or ask about a plan with higher throughput
503 No worker or cloud fallback can serve that specific model right now Rare, and fails fast on purpose rather than hanging — retry, or pick a more broadly available model
Response looks truncated max_tokens set too low, or a reasoning model spent its budget “thinking” before answering Raise max_tokens; reasoning-style cloud models need generous limits (128+)

What’s next?