VeeVECTOREE
SKILL · CLI · GATEWAY

Give your
agent a
backend.

Everything an AI agent needs to ship — models · templates · tools · auth · database · storage One cloud, from demo to production.

// $5 AFTER CLI LINK · PAYG WALLET · NO MONTHLY FEE

Vee, the Vectoree mascotVEE · YOUR CLOUD FOREMAN
$ set up https:
One key · every providerxAIOpenAIGoogleAnthropicDeepSeekOpenAI-compatible · one project key

Get started.

4 STEPS

// login, link, pick a model, then call the gateway or create a table

01
Install the skill
set up https://vectoree.cn/SKILL.md
02
Login & link
npx @vectoree/cli login --api-url https://vectoree.cn
03
Pick a model
vectoree/auto · half the price of Flash, all of Flash's intelligence
04
Call the API
gateway HTTP · db create
SERVER.TSOPENAI SDK · COMPATIBLE
import OpenAI from "openai"

const client = new OpenAI({
  baseURL: "https://vectoree.cn/api/v1",
  apiKey: process.env.VECTOREE_API_KEY,
});

// one project key. OpenAI-compatible.
const completion = await client.chat.completions.create({
  model: "vectoree/auto",
  messages: [{ role: "user",
    content: "Hello from Vectoree" }],
});

// catalog model — start with vectoree/auto

Skill + CLI.
OpenAI-compatible.

The coding agent installs the skill and runs atomic CLI commands.

set up https://vectoree.cn/SKILL.md
01 Install: set up https://vectoree.cn/SKILL.md
02 CLI: login · link · db · storage · ai
03 App runtime: OpenAI SDK + Vectoree baseURL (server-side key)

Point Codex at Vectoree.

ONE COMMAND

// switches Codex CLI (or ChatGPT desktop) to the Model Gateway

macOS / Linux
bash <(curl -fsSL https://vectoree.cn/scripts/codex-vectoree-setup.sh)
Windows PowerShell
irm https://vectoree.cn/scripts/codex-vectoree-setup.ps1 | iex
Agent prompt
Paste into Cursor, Claude Code, or any coding agent
Point my Codex CLI at Vectoree as the model provider.
Run the one-click setup (macOS/Linux):
  bash <(curl -fsSL https://vectoree.cn/scripts/codex-vectoree-setup.sh)
Or on Windows PowerShell:
  irm https://vectoree.cn/scripts/codex-vectoree-setup.ps1 | iex
Prefer menu 1 (project key) or 2 (employee key) so the script opens a browser and mints the key (same device-code flow as vectoree login).
Follow C07 in the Vectoree skill if the script is unavailable:
https://raw.githubusercontent.com/VectoreeAI/vectoree-skills/main/skills/vectoree/scenarios/model-gateway.md
Full guide →

Pay for tokens.
No monthly fee.

PAYG

// $5 welcome credit after CLI login + link · then top up the org wallet · billed per model

START
$0
per month · no subscription
  • ■Sign up and create a project
  • ■$5 welcome credit after CLI login + link
  • ■Agent skill + @vectoree/cli
  • ■Auth, Database, Storage included
GET STARTED
HOW YOU PAY
PAYG
WALLET
org top-up · billed per model
  • ■Stripe top-up into the org wallet
  • ■Gateway calls deduct catalog prices
  • ■Empty wallet cannot call the gateway
  • ■Usage in Organization → Billing
OPEN CONSOLE →

Before you
start.

FAQ

// pick a template · paste the prompt · the agent takes over

What is Vectoree?

Vectoree is a backend for coding agents and the apps they ship. Model calls, user login, Postgres, and object storage sit in one console, on one project key. The agent connects the project. Your app keeps calling it the way it already does.

Why do coding agents need it?

A coding agent can already write the UI and the business logic. Models, login, the database, and files are often still spread across separate services and separate keys. Vectoree turns those into one backend the agent can bind. Bind the project once. The app uses that one key.

What does Vectoree provide?

One console includes Model Plaza, Tool Hub, Auth, Postgres, and object storage. Model Plaza calls many models with one project key. Tool Hub gives the agent tools it can call. Auth is email and Google sign-in for your app users. Postgres holds tables and data. Object storage holds avatars, assets, and user files.

Who is Vectoree for?

Three kinds of people can use it to get their work done. A startup team or organization can keep several projects under one org: each project is built on its own, while members, billing, and settings are managed together, and usage draws from the organization wallet. An individual developer who already has an app keeps their code and the OpenAI SDK, and connects models, login, the database, and storage to Vectoree. Someone with no technical background who wants to build their own agent can start from a console template, hand the prompt to Cursor, Claude Code, or Codex, and let the agent connect Vectoree, then keep building on that base.

How does my agent use it?

Install the Vectoree Skill in Cursor, Claude Code, or Codex, then login and link. The agent binds the current project to Vectoree. After that it can create tables, upload files, and look up models through the Skill. Your app still uses the OpenAI SDK, pointed at Vectoree with the project key.

How does it charge?

Sign-up and the basics are free. There is no monthly fee. After you link a project you get $5, enough to try models and Tool Hub. Later usage bills from the organization wallet. An empty wallet stops the gateway. Auth, the database, and storage come with the project.

Can I keep existing code?

Yes. Vectoree follows the OpenAI endpoint, so you keep your SDK, business logic, and prompts. Point the API base URL at Vectoree and switch the key, and model calls move over. The code stays in your repo. Once the agent binds that project, login, the database, and storage can be used with it.

Vee
Vee
Vee
“Vee is watching your tokens.”
— EVERY CLOUD FOREMAN, EVER

Install the skill.
Link a project.

YOUR AGENT CAN LOGIN AND LINK BEFORE THIS PAGE FINISHES SCROLLING.

GET STARTED →