Build with Verge.

Your first request

Set up your account, choose a model, and send a message.

1. 1Open your consoleSign in and check that your account has available credit.
2. 2Create an API keySet its quota and permissions. Save it on your server as VERGE_API_KEY.
3. 3Choose a modelUse the exact model ID from the live catalogue in your request.

curl https://vergeapi.com/v1/chat/completions \
  -H "Authorization: Bearer $VERGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4.1-mini",
    "messages": [{"role": "user", "content": "Hello, Verge!"}],
    "max_tokens": 128
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://vergeapi.com/v1",
    api_key=os.environ["VERGE_API_KEY"],
)

response = client.chat.completions.create(
    model="openai/gpt-4.1-mini",
    messages=[{"role": "user", "content": "Hello, Verge!"}],
    max_tokens=128,
)

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

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://vergeapi.com/v1",
  apiKey: process.env.VERGE_API_KEY,
});

const response = await client.chat.completions.create({
  model: "openai/gpt-4.1-mini",
  messages: [{ role: "user", content: "Hello, Verge!" }],
  max_tokens: 128,
});

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

Authentication

Send your Verge API key in the Authorization header with every request. Use a separate key for each application so you can manage its quota, expiry, and model permissions independently.

Authorization: Bearer $VERGE_API_KEY

Create and manage keys in your API keys page. Your account balance and the key's limits both apply to API requests.

Chat completions

Send a conversation as a messages array to POST /v1/chat/completions. Select a model with its exact catalogue ID. The first-request example above uses openai/gpt-4.1-mini.

Read the generated text from choices[0].message.content. Image input, tools, and other model-specific capabilities depend on the model you select.

Stream a response

Add stream: true to receive the response as it is generated. With the Python client from the first example, read the content in each chunk.

stream = client.chat.completions.create(
    model="openai/gpt-4.1-mini",
    messages=[{"role": "user", "content": "Hello, Verge!"}],
    stream=True,
    max_tokens=128,
)

for chunk in stream:
    if chunk.choices:
        print(chunk.choices[0].delta.content or "", end="")

For a direct HTTP integration, check the response status before reading the stream. Parse server-sent events until data: [DONE].

Create embeddings

Choose an embedding model and send the text as input. Embeddings use a separate endpoint from chat completions.

curl https://vergeapi.com/v1/embeddings \
  -H "Authorization: Bearer $VERGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/text-embedding-3-small",
    "input": "Text to embed"
  }'

Models and pricing

The live model catalogue shows available models, supported capabilities, and current prices. Use GET /v1/models to list the models available to your key.

curl https://vergeapi.com/v1/models \
  -H "Authorization: Bearer $VERGE_API_KEY"

Catalogue pricing is shown in USD. Charges can vary by input, output, cached tokens, and context length. Check a model's details before using provider-specific tools or capabilities.

Models with dynamic routing prices are not enabled until their charges can be settled accurately.

Errors and retries

Check the HTTP status and response body together. The error message explains the specific reason the request failed.