Let users bring their AI wallet to your app.

Stop eating your users' LLM bills or building complex billing meters. Users connect their funded Zorveus wallet to your AI app with custom spending caps. You execute model inference, Zorveus handles the billing.

5-minute setup · TypeScript & Python SDKs · No credit card required

Connect user wallets in minutes

Allow end-users to authorize your app with OAuth 2.0 PKCE and fund model inference directly.

1-initiate-auth.ts

import { ZorveusOAuth } from "@zorveus/sdk";

// 1. Generate PKCE verifier, S256 challenge, and state CSRF token
const pkce = ZorveusOAuth.generatePKCE();

sessionStorage.setItem("zorveus_code_verifier", pkce.codeVerifier);
sessionStorage.setItem("zorveus_auth_state", pkce.state);

// 2. Build authorization URL (requests inference & model access)
const authUrl = ZorveusOAuth.getAuthorizationUrl({
  clientId: process.env.NEXT_PUBLIC_ZORVEUS_CLIENT_ID!,
  redirectUri: "https://myapp.com/oauth/callback",
  state: pkce.state,
  codeChallenge: pkce.codeChallenge,
  scopes: ["inference:write", "models:*"],
});

// 3. Redirect browser to Zorveus authorization consent page
window.location.href = authUrl;

import { ZorveusOAuth } from "@zorveus/sdk";

// 1. Generate PKCE verifier, S256 challenge, and state CSRF token
const pkce = ZorveusOAuth.generatePKCE();

sessionStorage.setItem("zorveus_code_verifier", pkce.codeVerifier);
sessionStorage.setItem("zorveus_auth_state", pkce.state);

// 2. Build authorization URL (requests inference & model access)
const authUrl = ZorveusOAuth.getAuthorizationUrl({
  clientId: process.env.NEXT_PUBLIC_ZORVEUS_CLIENT_ID!,
  redirectUri: "https://myapp.com/oauth/callback",
  state: pkce.state,
  codeChallenge: pkce.codeChallenge,
  scopes: ["inference:write", "models:*"],
});

// 3. Redirect browser to Zorveus authorization consent page
window.location.href = authUrl;

1_initiate_auth.py

from zorveus import ZorveusOAuth

# 1. Generate PKCE parameters for OAuth flow
pkce = ZorveusOAuth.generate_pkce()
session["code_verifier"] = pkce.code_verifier
session["oauth_state"] = pkce.state

# 2. Build authorization URL
auth_url = ZorveusOAuth.get_authorization_url(
    client_id="YOUR_CLIENT_ID",
    redirect_uri="https://myapp.com/oauth/callback",
    state=pkce.state,
    code_challenge=pkce.code_challenge,
    scopes=["inference:write", "models:*"],
)

# 3. Redirect user browser to auth_url

from zorveus import ZorveusOAuth

# 1. Generate PKCE parameters for OAuth flow
pkce = ZorveusOAuth.generate_pkce()
session["code_verifier"] = pkce.code_verifier
session["oauth_state"] = pkce.state

# 2. Build authorization URL
auth_url = ZorveusOAuth.get_authorization_url(
    client_id="YOUR_CLIENT_ID",
    redirect_uri="https://myapp.com/oauth/callback",
    state=pkce.state,
    code_challenge=pkce.code_challenge,
    scopes=["inference:write", "models:*"],
)

# 3. Redirect user browser to auth_url

Terminal: 1. OAuth Redirect

# 1. Generate PKCE verifier & S256 challenge
# 2. Redirect user browser to authorization URL:
https://api.zorveus.com/oauth/authorize\
  ?client_id=YOUR_CLIENT_ID\
  &redirect_uri=https://myapp.com/oauth/callback\
  &response_type=code\
  &scope=inference:write%20models:*\
  &code_challenge=CODE_CHALLENGE\
  &code_challenge_method=S256

# 1. Generate PKCE verifier & S256 challenge
# 2. Redirect user browser to authorization URL:
https://api.zorveus.com/oauth/authorize\
  ?client_id=YOUR_CLIENT_ID\
  &redirect_uri=https://myapp.com/oauth/callback\
  &response_type=code\
  &scope=inference:write%20models:*\
  &code_challenge=CODE_CHALLENGE\
  &code_challenge_method=S256

2-exchange-token.ts

import { ZorveusOAuth } from "@zorveus/sdk";

// 1. Read stored state & verifier on callback page
const savedState = sessionStorage.getItem("zorveus_auth_state");
const codeVerifier = sessionStorage.getItem("zorveus_code_verifier");

// 2. Validate callback state CSRF token
const validation = ZorveusOAuth.validateCallback(window.location.href, savedState);

// 3. Exchange authorization code for user access token
const tokenResponse = await ZorveusOAuth.exchangeToken({
  clientId: process.env.NEXT_PUBLIC_ZORVEUS_CLIENT_ID!,
  code: validation.code!,
  codeVerifier: codeVerifier!,
  redirectUri: "https://myapp.com/oauth/callback",
});

// Note: The OAuth access_token IS the user's API key for inference
const userApiKey = tokenResponse.access_token;

import { ZorveusOAuth } from "@zorveus/sdk";

// 1. Read stored state & verifier on callback page
const savedState = sessionStorage.getItem("zorveus_auth_state");
const codeVerifier = sessionStorage.getItem("zorveus_code_verifier");

// 2. Validate callback state CSRF token
const validation = ZorveusOAuth.validateCallback(window.location.href, savedState);

// 3. Exchange authorization code for user access token
const tokenResponse = await ZorveusOAuth.exchangeToken({
  clientId: process.env.NEXT_PUBLIC_ZORVEUS_CLIENT_ID!,
  code: validation.code!,
  codeVerifier: codeVerifier!,
  redirectUri: "https://myapp.com/oauth/callback",
});

// Note: The OAuth access_token IS the user's API key for inference
const userApiKey = tokenResponse.access_token;

2_exchange_token.py

from zorveus import ZorveusOAuth

# Read query parameters and saved session state
saved_verifier = session.get("code_verifier")

# Exchange authorization code for user access token
token_data = ZorveusOAuth.exchange_token(
    client_id="YOUR_CLIENT_ID",
    code=auth_code,
    code_verifier=saved_verifier,
    redirect_uri="https://myapp.com/oauth/callback",
)

# Note: The OAuth access_token IS the user's API key for inference
user_api_key = token_data.access_token

from zorveus import ZorveusOAuth

# Read query parameters and saved session state
saved_verifier = session.get("code_verifier")

# Exchange authorization code for user access token
token_data = ZorveusOAuth.exchange_token(
    client_id="YOUR_CLIENT_ID",
    code=auth_code,
    code_verifier=saved_verifier,
    redirect_uri="https://myapp.com/oauth/callback",
)

# Note: The OAuth access_token IS the user's API key for inference
user_api_key = token_data.access_token

Terminal: 2. Token Exchange

# Exchange authorization code for user access token
curl -X POST https://api.zorveus.com/oauth/token \
  -H "Content-Type: application/json" \
  -d '{
    "client_id": "YOUR_CLIENT_ID",
    "grant_type": "authorization_code",
    "code": "AUTH_CODE_FROM_CALLBACK",
    "code_verifier": "PKCE_CODE_VERIFIER",
    "redirect_uri": "https://myapp.com/oauth/callback"
  }'

# Exchange authorization code for user access token
curl -X POST https://api.zorveus.com/oauth/token \
  -H "Content-Type: application/json" \
  -d '{
    "client_id": "YOUR_CLIENT_ID",
    "grant_type": "authorization_code",
    "code": "AUTH_CODE_FROM_CALLBACK",
    "code_verifier": "PKCE_CODE_VERIFIER",
    "redirect_uri": "https://myapp.com/oauth/callback"
  }'

3-run-inference-zorveus.ts

import { Zorveus } from "@zorveus/sdk";

// 1. Initialize Zorveus client using the user's access_token as the API key
const client = new Zorveus({ apiKey: userApiKey });

// 2. Execute AI inference charged directly to the user's wallet
const response = await client.chat.completions.create({
  model: "openai/gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

import { Zorveus } from "@zorveus/sdk";

// 1. Initialize Zorveus client using the user's access_token as the API key
const client = new Zorveus({ apiKey: userApiKey });

// 2. Execute AI inference charged directly to the user's wallet
const response = await client.chat.completions.create({
  model: "openai/gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

3-run-inference-openai-sdk.ts

import OpenAI from "openai";

// 1. Point Official OpenAI SDK to Zorveus Edge gateway proxy
const openai = new OpenAI({
  baseURL: "https://api.zorveus.com/v1",
  apiKey: userApiKey, // OAuth access_token acts as the API key
});

// 2. Standard OpenAI API call billed directly to user's wallet
const response = await openai.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

import OpenAI from "openai";

// 1. Point Official OpenAI SDK to Zorveus Edge gateway proxy
const openai = new OpenAI({
  baseURL: "https://api.zorveus.com/v1",
  apiKey: userApiKey, // OAuth access_token acts as the API key
});

// 2. Standard OpenAI API call billed directly to user's wallet
const response = await openai.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

3-run-inference-zorveus-openai.ts

import { ZorveusOpenAI } from "@zorveus/sdk/openai";

// 1. Initialize ZorveusOpenAI adapter with user access_token as the API key
const client = new ZorveusOpenAI({ apiKey: userApiKey });

// 2. Standard OpenAI API call billed directly to user's wallet
const response = await client.chat.completions.create({
  model: "openai/gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

import { ZorveusOpenAI } from "@zorveus/sdk/openai";

// 1. Initialize ZorveusOpenAI adapter with user access_token as the API key
const client = new ZorveusOpenAI({ apiKey: userApiKey });

// 2. Standard OpenAI API call billed directly to user's wallet
const response = await client.chat.completions.create({
  model: "openai/gpt-4o",
  messages: [{ role: "user", content: "Summarize Q3 financial report." }],
});

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

3-run-inference-vercel-ai.ts

import { createOpenAI } from "@ai-sdk/openai";
import { generateText } from "ai";

// 1. Create OpenAI provider pointing to Zorveus Edge gateway
const zorveus = createOpenAI({
  baseURL: "https://api.zorveus.com/v1",
  apiKey: userApiKey, // OAuth access_token acts as the API key
});

// 2. Generate text using Vercel AI SDK billed to user wallet
const { text } = await generateText({
  model: zorveus("openai/gpt-4o"),
  prompt: "Summarize Q3 financial report.",
});

console.log(text);

import { createOpenAI } from "@ai-sdk/openai";
import { generateText } from "ai";

// 1. Create OpenAI provider pointing to Zorveus Edge gateway
const zorveus = createOpenAI({
  baseURL: "https://api.zorveus.com/v1",
  apiKey: userApiKey, // OAuth access_token acts as the API key
});

// 2. Generate text using Vercel AI SDK billed to user wallet
const { text } = await generateText({
  model: zorveus("openai/gpt-4o"),
  prompt: "Summarize Q3 financial report.",
});

console.log(text);

3_run_inference_zorveus.py

from zorveus import Zorveus

# 1. Instantiate client with user access_token as the API key
client = Zorveus(api_key=user_api_key)

# 2. Execute AI inference charged directly to user wallet
response = client.chat.completions.create(
    model="openai/gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

from zorveus import Zorveus

# 1. Instantiate client with user access_token as the API key
client = Zorveus(api_key=user_api_key)

# 2. Execute AI inference charged directly to user wallet
response = client.chat.completions.create(
    model="openai/gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

3_run_inference_openai_sdk.py

from openai import OpenAI

# 1. Point official OpenAI SDK to Zorveus Edge gateway proxy
client = OpenAI(
    base_url="https://api.zorveus.com/v1",
    api_key=user_api_key,  # OAuth access_token acts as the API key
)

# 2. Standard OpenAI API call billed directly to user wallet
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

from openai import OpenAI

# 1. Point official OpenAI SDK to Zorveus Edge gateway proxy
client = OpenAI(
    base_url="https://api.zorveus.com/v1",
    api_key=user_api_key,  # OAuth access_token acts as the API key
)

# 2. Standard OpenAI API call billed directly to user wallet
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

3_run_inference_zorveus_openai.py

from zorveus.openai import ZorveusOpenAI

# 1. Instantiate OpenAI SDK adapter with user access_token as the API key
client = ZorveusOpenAI(api_key=user_api_key)

# 2. Standard OpenAI API call billed directly to user wallet
response = client.chat.completions.create(
    model="openai/gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

from zorveus.openai import ZorveusOpenAI

# 1. Instantiate OpenAI SDK adapter with user access_token as the API key
client = ZorveusOpenAI(api_key=user_api_key)

# 2. Standard OpenAI API call billed directly to user wallet
response = client.chat.completions.create(
    model="openai/gpt-4o",
    messages=[{"role": "user", "content": "Summarize Q3 financial report."}],
)

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

Terminal: 3. Inference Request

# Execute model completion using user access_token as Bearer token
curl -X POST https://api.zorveus.com/v1/chat/completions \
  -H "Authorization: Bearer zrv_usr_live_9f8e7d6c5b" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o",
    "messages": [
      {"role": "user", "content": "Summarize Q3 financial report."}
    ]
  }'

# Execute model completion using user access_token as Bearer token
curl -X POST https://api.zorveus.com/v1/chat/completions \
  -H "Authorization: Bearer zrv_usr_live_9f8e7d6c5b" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o",
    "messages": [
      {"role": "user", "content": "Summarize Q3 financial report."}
    ]
  }'

npm i @zorveus/sdk

View on npm React bindings GitHub

Your app makes requests.
Zorveus handles billing.

Users connect their Zorveus wallet through OAuth. Your app uses their access token to request models, and Zorveus charges their wallet for usage.

Your app

Sends a model request with the user's access token.

Zorveus

Checks access and spending limits, then routes the request.

Model providers

OpenAI, Anthropic, Gemini, and other supported providers.

Zorveus bills usage toUsage billed to

User's Zorveus wallet

The user approves how much your app can spend.

User-funded requests: Usage is charged to the connected wallet. Your app does not collect payment details.
User-controlled limits: Users approve how much your app can spend when they connect.
One model gateway: Use the same access token to call supported models across providers.

Start building in under 5 minutes

Create a developer workspace, register your client application, and connect your first user wallet.

Create developer account Explore on GitHub

Let users bring their AI wallet to your app.

Connect user wallets in minutes

Your app makes requests. Zorveus handles billing.

Start building in under 5 minutes

Your app makes requests.
Zorveus handles billing.