Quickstart: Developers

Sign up, add as little as $1 in credits, create an API key, and send a request. Most of this is account setup. The request itself is one command, and it goes straight to the model you name. Before your first request you need:

A Gloo AI Studio account and organization.
Funded prepaid credits (Inference is pay as you go).
An API key.

Once you have a key, this is the whole request:

curl https://platform.ai.gloo.com/ai/v2/direct/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOO_API_KEY" \
  -d '{
    "model": "gloo-anthropic-claude-opus-5.5",
    "input": [{"role": "user", "content": "Explain prompt caching in two sentences."}]
  }'

This is a direct endpoint: no platform fee, and you pay the model’s published rate. You choose the model, and Gloo adds nothing between your prompt and the provider. Direct is the default for this quickstart and the simplest place to start. Nothing screens your input or reviews the output, so your application owns its own safety. To have Gloo apply guardrails and output moderation, use the Guarded endpoint instead: same request, different path.

Inference needs funded credits. Grounded endpoints and Search also need Data Engine access, which is separate and currently at capacity for new organizations. See Billing & Plans.

The steps below walk through each prerequisite.

Step 1: Create Your Account & Organization

Start by creating your account at Gloo AI Studio. Once your email is verified, the onboarding wizard will guide you through setting up your organization:

Create your Organization: Give your team a name to manage projects and collaboration.
Invite your Team: (Optional) Add email addresses to invite collaborators immediately, or skip this step.

As the account creator, you are the Organization Admin. You can invite team members, create Publishers, and manage your organization at any time from the Studio sidebar. See the Admin Quickstart for a full walkthrough.

Step 2: Choose Your Products & Add Credits

Onboarding asks what brings you to Gloo. For API access, choose Inference. It’s pay as you go, funded by prepaid credits. You can add Gloo Code in the same checkout if you want the coding agent too.

Add opening credits: Pick a starting credit balance on the review screen: $10, $25, $50 or $100, or any custom amount from $1 to $100. API usage draws it down at each model’s rate, and you can turn on auto-refill later from the Billing page.
Turn on auto-refill: From the Billing page, auto-refill charges your saved card when your balance runs low, so you do not have to watch the balance. See Prepaid Credits.
Pay once for everything: One checkout covers your opening credits and the first month of any Gloo Code tier. Your card is saved for future fills.
Know the cap: Credit-funded API usage is capped per organization at $4,000 per calendar month by default, regardless of balance. Support can raise or lower it on request.

Once your products are active, continue below to create an API key. After your key is created, you can either call the API directly or open the Playground in Studio to chat with models in the browser. The Playground asks you to paste your key, and both routes count against your usage limit.

Step 3: Create Your API Key

The Gloo AI API uses API keys for authentication. A single API key is all you need to make API calls.

Creating an API key requires a funded wallet or a payment method on file. If you skipped adding credits in Step 2, the API Keys page sends you to the Billing page first. See Prepaid Credits to add credits.

Stuck creating a key? See Get help. Here’s how to get your API key:

Click API Keys in the Studio sidebar.
Click Create New Key. Enter a name for the key and click Create.
Copy the API key immediately. It is shown only once; after that, Studio displays only the obfuscated value (e.g., sk_...3456).

Treat your API key like a password. Store it in a secure location, such as a secrets manager or environment variable. Do not commit it to source control.

Step 4: Set Your API Key

Set GLOO_API_KEY to the key you copied in Step 3. For example, in your shell:

export GLOO_API_KEY="sk_your_api_key_here"

Or add it to a .env file in your project:

GLOO_API_KEY="sk_your_api_key_here"

Step 5: Make An API Call

Send your API key as a bearer token in the Authorization header. Pick the request format you already know. Both are direct endpoints, and both are well tested.

Direct Responses
Direct Completions

The recommended format for new builds, with text, vision, tools and image generation.

import os
import requests

response = requests.post(
    "https://platform.ai.gloo.com/ai/v2/direct/responses",
    headers={
        "Content-Type": "application/json",
        "Authorization": f"Bearer {os.getenv('GLOO_API_KEY')}"
    },
    json={
        "model": "gloo-anthropic-claude-opus-5.5",
        "input": [
            {"role": "user", "content": "Explain prompt caching in two sentences."}
        ]
    }
)

# Find the assistant message in the typed output array (a reasoning item
# may precede it on reasoning-capable models).
output = response.json()["output"]
message = next(item for item in output if item["type"] == "message")
print(message["content"][0]["text"])

curl https://platform.ai.gloo.com/ai/v2/direct/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOO_API_KEY" \
  -d '{
    "model": "gloo-anthropic-claude-opus-5.5",
    "input": [
      {
        "role": "user",
        "content": "Explain prompt caching in two sentences."
      }
    ]
  }'

Learn more in Direct Responses.

The OpenAI Chat Completions request shape. Use it when you already have tooling built on messages.

import os
import requests

response = requests.post(
    "https://platform.ai.gloo.com/ai/v2/direct/chat/completions",
    headers={
        "Content-Type": "application/json",
        "Authorization": f"Bearer {os.getenv('GLOO_API_KEY')}"
    },
    json={
        "model": "gloo-anthropic-claude-opus-5.5",
        "messages": [
            {"role": "user", "content": "Explain prompt caching in two sentences."}
        ]
    }
)

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

curl https://platform.ai.gloo.com/ai/v2/direct/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOO_API_KEY" \
  -d '{
    "model": "gloo-anthropic-claude-opus-5.5",
    "messages": [
      {"role": "user", "content": "Explain prompt caching in two sentences."}
    ]
  }'

Learn more in Direct Completions.

Request failed? The usual causes are a mistyped key, an empty balance, or a wrong model ID. See Get help. A Direct Responses reply looks like this (abridged):

{
  "object": "response",
  "model": "gloo-anthropic-claude-opus-5.5",
  "output": [
    {
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        {
          "type": "output_text",
          "text": "Prompt caching stores ..."
        }
      ]
    }
  ],
  "usage": {
    "input_tokens": 14,
    "output_tokens": 52,
    "total_tokens": 66
  }
}

Step 6: Keep your wallet flowing

Without auto-refill, requests stop when your balance reaches zero. Turn it on once and Gloo keeps your wallet topped up, within the limits below.

On the Billing page, find the Auto-refill switch on the Wallet card.
Set the amount to add each time ($10 to $100).
Flip the switch on and confirm.

Gloo then charges your saved card when your balance runs low, up to $1,000 per rolling 7 days. If you reach that cap, or your organization’s monthly API spend limit, requests can still stop. Only organization Admins can add credits. See Prepaid Credits.

Want Gloo Code too? It is a monthly plan per seat, from $20 a month, and it does not draw from your credits. See Pricing.

Choose your path

Direct is the default, and you can change your mind later. The request body stays the same when you switch between direct and guarded Responses, so moving is a URL change.

Direct Responses

Start here for new builds. You name the model and your request goes straight to it. No platform fee.

Direct Completions

Start here if you have Chat Completions tooling. Same directness, the messages format you already use. Well tested and rigorous. No platform fee.

Guarded

Choose this when Gloo should handle safety. Input guardrails, output moderation and tradition settings, and with Completions also model_family and auto_routing. Adds a 6.5% platform fee. See below.

Grounded

Answer from your own content. Retrieval over your library, with optional citations. Needs Data Engine access.

Not sure? Endpoint Types compares them side by side.

Guarded endpoints: when and how

Direct sends your request straight to the model you name. Guarded runs Gloo’s safety and values pipeline around the same request. What Guarded adds:

Input guardrails and output moderation. Requests that guardrails reject outright return a 403 error, and moderation can replace a response.
Values-aligned responses through the tradition parameter.
On Completions, model_family selection and opt-in auto_routing.
A 6.5% platform fee on top of the model’s published rate. Direct endpoints add no fee.

When to choose it: pick Guarded when any part of your traffic needs screening, or when you want values-aligned (tradition) responses, and you would rather Gloo apply that than build it yourself. Stay on Direct when your application owns its own safety and prompting. How to switch: keep the request body and change the path. Only the URL differs.

# Direct
curl https://platform.ai.gloo.com/ai/v2/direct/responses ...

# Guarded: same request, different path
curl https://platform.ai.gloo.com/ai/v2/guarded/responses ...

# Direct
curl https://platform.ai.gloo.com/ai/v2/direct/chat/completions ...

# Guarded: same request, different path
curl https://platform.ai.gloo.com/ai/v2/guarded/chat/completions ...

If you use an OpenAI SDK, change the base URL from https://platform.ai.gloo.com/ai/v2/direct to https://platform.ai.gloo.com/ai/v2/guarded. Learn more in Guarded Responses, Guarded Completions and Endpoint Types.

Coming from another provider?

If you already call OpenAI or another gateway, you can usually move with three changes. See Migrate from another provider for the step-by-step version.

Point your client at https://platform.ai.gloo.com/ai/v2/direct.
Use your Gloo API key as the key.
Use a Gloo model ID, such as gloo-anthropic-claude-opus-5.5.

What to do next

Keep it running: wallet and auto-refill

Turn on auto-refill so Gloo tops up your wallet when your balance runs low.

Add Gloo Code seats

A coding agent for you and your team, from $20 per seat per month. It signs in with your Gloo AI account and is billed separately from your credits. See Pricing.

Try tools

Let the model call your functions and return structured results.

Stream responses

Show tokens as they arrive.

Cut repeat costs

Reuse prompt prefixes at a discounted rate.

Use the OpenAI SDK

Python and Node.js with Gloo as the base URL.

Explore models

More than 160 models, with context windows and rates.

Ground answers in your content

Ingest your library and answer from it. Needs Data Engine access.

Getting help

Stuck? Check the common problems, then open Support at the bottom of the Studio sidebar. It is the fastest way to reach us, for API, account, billing and key questions. See Get help.

Start Here

Inference and APIs

Data Engine

Studio

Gloo Code

Learn More

Legacy

Quickstart: Developers

Step 1: Create Your Account & Organization

Step 2: Choose Your Products & Add Credits

Step 3: Create Your API Key

Step 4: Set Your API Key

Step 5: Make An API Call

Step 6: Keep your wallet flowing

Choose your path

Direct Responses

Direct Completions

Guarded

Grounded

Guarded endpoints: when and how

Coming from another provider?

What to do next

Keep it running: wallet and auto-refill

Add Gloo Code seats

Try tools

Stream responses

Cut repeat costs

Use the OpenAI SDK

Explore models

Ground answers in your content

Getting help

​Step 1: Create Your Account & Organization

​Step 2: Choose Your Products & Add Credits

​Step 3: Create Your API Key

​Step 4: Set Your API Key

​Step 5: Make An API Call

​Step 6: Keep your wallet flowing

​Choose your path

Direct Responses

Direct Completions

Guarded

Grounded

​Guarded endpoints: when and how

​Coming from another provider?

​What to do next

Keep it running: wallet and auto-refill

Add Gloo Code seats

Try tools

Stream responses

Cut repeat costs

Use the OpenAI SDK

Explore models

Ground answers in your content

​Getting help

Step 1: Create Your Account & Organization

Step 2: Choose Your Products & Add Credits

Step 3: Create Your API Key

Step 4: Set Your API Key

Step 5: Make An API Call

Step 6: Keep your wallet flowing

Choose your path

Guarded endpoints: when and how

Coming from another provider?

What to do next

Getting help