Documentation

GloopAPI integration guide

Integration overview

Get a key, discover models, and send your first request.

GloopAPI provides models, generation tasks, tools, and result files through a unified API. First create a dedicated key for your application, then discover the catalog available to that key. Models or tools shown on the website do not mean your key is authorized to call them.

Your first request

Create a key in your account's developer settings and put it in a server environment variable. GLOOP_API_BASE is the deployment's API origin without /v1; web examples use the currently configured API address.

export GLOOP_API_BASE="https://api.gloopapi.com"
curl "$GLOOP_API_BASE/v1/models" \
  -H "Authorization: Bearer $GLOOP_API_KEY"

Choose a text-capable model from the returned catalog, save it as GLOOP_MODEL, then send a minimal text request:

curl "$GLOOP_API_BASE/v1/chat/completions" \
  -H "Authorization: Bearer $GLOOP_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"model\":\"$GLOOP_MODEL\",\"messages\":[{\"role\":\"user\",\"content\":\"你好\"}]}"

These examples explain how to use the API. Available models and prices are determined by the current catalog and quotes. Reading or copying documentation does not send requests.

Choose a workflow

Text generation is suited to consuming responses directly. For persistent images and videos, first read the capability catalog, prepare inputs, obtain a quote, and confirm creation, then retain the request key for recovery. Synchronous images and speech return results directly; persistent task recovery endpoints do not apply to them.

Start with Authentication, Model selection, and SDK integration. For streaming reads, see Streaming responses; for task recovery, see Asynchronous tasks and idempotency.

API entry points