#Quickstart

Make your first request to GOB-5.5 in under two minutes.

##1. Get an API key

text

gob-live-x4f9k3m2nQ8wHvL7pYrJ5tA1cBdE6fG2sK9oXz0aPq

Treat your API key like a password. Don't commit it to git, don't paste it in chat, don't put it in client-side code.

##2. Install the SDK (optional)

We ship official SDKs for Python and Node.js, but the API is plain HTTP — you can use curl, fetch, requests, or any HTTP client.

bash

# Python
pip install gpt-gob
# Node.js / TypeScript
npm install @gpt-gob/sdk

##3. Make your first call

###cURL

bash

curl https://api.gpt-gob.ai/v1/chat/completions \
  -H "Authorization: Bearer $GOB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gob-5.5",
    "messages": [
      {"role": "user", "content": "explain recursion in one sentence"}
    ]
  }'

###Python

python

from gpt_gob import GPTGob

client = GPTGob(api_key="gob-...")

response = client.chat.completions.create(
    model="gob-5.5",
    messages=[
        {"role": "user", "content": "explain recursion in one sentence"}
    ],
)

print(response.choices[0].message.content)

###Node.js

javascript

import { GPTGob } from "@gpt-gob/sdk";

const client = new GPTGob({ apiKey: process.env.GOB_API_KEY });

const response = await client.chat.completions.create({
  model: "gob-5.5",
  messages: [
    { role: "user", content: "explain recursion in one sentence" },
  ],
});

console.log(response.choices[0].message.content);

##4. What you get back

json

{
  "id": "gob-cmpl-9F2kQxLm7HpVbNa3",
  "object": "chat.completion",
  "created": 1746998400,
  "model": "gob-5.5",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "recursion is when a goblin sends another goblin to do the same job, and that one sends another, until the smallest goblin in the cave finds the answer and passes it back up the chain."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 38,
    "total_tokens": 47,
    "grs_score": 0.91
  }
}

Notice the extra grs_score field in usage — that's the Goblin Reward Signal score for the completion (0.0–1.0). The higher, the more goblin-aligned the output. See Goblin Reward Signal for details.

##Next steps

▸Streaming — get token-by-token responses
▸Models — pick the right tier (scout, base, deep, horde)
▸Parameters — tune mining_depth, shadow_attention, etc.