Technical Documentation

Unified API reference + integration guides for every model on we64.com. OpenAI + Anthropic-compatible endpoints — drop-in replacement for the official openai SDK. Works with Cursor, Claude Code, IntelliJ, N8N, Dify, MCP clients, and more.

Jump to:Base URL & Auth MCP Python SDK Node.js cURL Cursor Claude Code IntelliJ IDEA Open Code Openclaw N8N Dify ⚠ Data & Hosting

Base URL & Authentication

All we64.com models use the same OpenAI-compatible endpoint.

Base URL

https://api.we64.com/v1

For tools that expect a trailing slash, use https://api.we64.com/v1/

API Key

we64_sk_...

Sent on WhatsApp + email after manual activation

Model name: use the exact model name provided in your activation WhatsApp message (e.g. deepseek-v4-pro, kimi-k3, glm-5-2). Custom model names are not supported — use the full slug from the model list.

Python · openai SDK

pip install openai from openai import OpenAI client = OpenAI( api_key="we64_sk_...", base_url="https://api.we64.com/v1"
) response = client.chat.completions.create( model="deepseek-v4-pro", messages=[ {"role": "user", "content": "Hello!"} ]
)
print(response.choices[0].message.content)

Node.js · openai SDK

npm install openai import OpenAI from "openai"; const openai = new OpenAI({ apiKey: "we64_sk_...", baseURL: "https://api.we64.com/v1",
}); async function main() { const completion = await openai.chat.completions.create({ model: "deepseek-v4-pro", messages: [{ role: "user", content: "Hello!" }], }); console.log(completion.choices[0].message);
} main();

cURL · terminal test

curl https://api.we64.com/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer we64_sk_..." \ -d '{ "model": "deepseek-v4-pro", "messages": [{"role": "user", "content": "Hello!"}] }'

MCP Endpoint

Use your models as tools in Claude Desktop, Cursor and any MCP client.

MCP URL

https://api.we64.com/v1/mcp

Streamable HTTP transport · JSON-RPC 2.0 · protocol version 2025-11-25

Auth

Authorization: Bearer we64_sk_...

Same API key as /v1/chat/completions · Bearer is enough — no OAuth.

Endpoint

/v1/mcp

Also accepted

/mcp · /v1/mcp/

Transport

Streamable HTTP

Protocol

JSON-RPC 2.0

Claude Desktop · settings.json

{ "mcpServers": { "we64-models": { "type": "http", "url": "https://api.we64.com/v1/mcp", "headers": { "Authorization": "Bearer we64_sk_your-api-key" } } }
}

Works in Claude Desktop / Cursor / any MCP HTTP client — set the URL + Bearer header and your models become tools you can call directly. Unlimited tokens, flat-rate.

Integration Guides

Plug we64.com into your tools

Step-by-step guides for popular AI tools, IDEs, and frameworks. All use the same OpenAI-compatible Base URL + API key.

Cursor

Enable custom OpenAI Base URL in Models settings · Pro account required

① Prerequisites: You need an active proxy connection. Your Cursor account must be Pro level or above to use third-party custom models.

② Configure Base URL: Open Cursor Settings (gear icon) → Models tab. Find the "OpenAI API Key" section, enable the custom switch, enter your Key, and expand "Override OpenAI Base URL" to fill in the address:

Override Base URL

https://api.we64.com/v1/

API Key

The key we send you on WhatsApp after activation

③ Select Models: Check the desired model in the model list, add a custom model name (must match our slug exactly — e.g. deepseek-v4-pro), click Add. Test the connection, save the settings, and confirm in the Chat session.

Claude Code (Local CLI)

Terminal AI programming assistant · Win/Mac/Linux · Anthropic-compatible

① System Requirements: macOS 10.15+ / Ubuntu 20.04+ / Windows 10+ (requires WSL or Git for Windows); 4GB+ RAM; Node.js 18+.

② Windows Install: Install Git for Windows first (default options: git-scm.com/install/windows), verify with git --version. Then install Claude Code in PowerShell:

irm https://claude.ai/install.ps1 | iex

③ macOS / Linux Install:

curl -fsSL https://claude.ai/install.sh | bash
# Or via Homebrew:
brew install --cask claude-code
# Or via NPM (requires Node 18+):
npm install -g @anthropic-ai/claude-code
# Verify:
claude --version

④ Configure settings.json: Config file location: Windows C:\Users\Username\.claude\settings.json; Mac/Linux ~/.claude/settings.json. Replace ANTHROPIC_AUTH_TOKEN with your API token and set ANTHROPIC_BASE_URL to our domain:

{ "env": { "ANTHROPIC_AUTH_TOKEN": "we64_sk_your-api-token", "ANTHROPIC_BASE_URL": "https://api.we64.com", "ANTHROPIC_MODEL": "deepseek-v4-pro", "API_TIMEOUT_MS": "3000000", "CLAUDE_CODE_MAX_OUTPUT_TOKENS": "200000" }
}

⑤ Start Using: Navigate to your project directory and run claude to start. On first run, follow the prompts to select a theme, confirm safety guidelines, and trust the directory.

IntelliJ IDEA

Install Claude Code GUI plugin · Add custom vendor via Vendor Management

① Install Plugin: Go to IDEA Settings → Plugins, search for "Claude Code GUI" (other code-related plugins work similarly) and install it.

② Access Vendor Management: Open the right sidebar → Plugins → Settings → Vendor Management. Select "Add Vendor" → Custom Configuration.

③ Fill Configuration: Vendor name can be customized, notes can be left blank. Fill in the following and click Save:

Request Address (Base URL)

https://api.we64.com (full server domain)

API Key

The key we send you on WhatsApp after activation

Model Mapping (ID)

Full model name from the model list

④ Enable and Chat: After saving, the configuration won't take effect automatically. You must manually click "Enable". Return to the chat interface, switch to this model, and start your conversation.

Open Code

Add custom provider · Fill URL / Key / full model name · Save

① Configuration: Add a custom provider with the following settings, then save to apply (leave request headers blank):

Provider ID

e.g., deepseek-v4-pro, kimi-k3, glm-5-2

Display Name

Custom name of your choice

URL

https://api.we64.com (or with /v1 suffix, test both)

API Secret

The key we send you on WhatsApp after activation

Model

Full model name from the model list (custom names not supported)

Openclaw

One-click deployment · Command-line wizard configuration · Win/Mac/Linux

① Installation: Chinese community version (recommended): clawd.org.cn; Official site: openclaw.ai. One-click installation in Windows Administrator PowerShell:

iwr -useb https://clawd.org.cn/install.ps1 | iex

② Enter Configuration Wizard: After installation, run openclaw onboard in the command line to enter configuration (skip download steps). Fill in Key / Base URL / Model ID as prompted.

③ Key Parameters: Fill in the following three items in the wizard:

API Key

Your we64.com API key

Base URL

https://api.we64.com/v1

Model ID

e.g., deepseek-v4-pro, kimi-k3, glm-5-2

N8N

Configure OpenAI-compatible provider in credentials · Call models in workflow nodes

① Configuration: Add an OpenAI-compatible provider credential: fill in the server domain + /v1 suffix as the request address, enter your Key, and save.

Base URL

https://api.we64.com/v1

API Key

The key we send you on WhatsApp after activation

Model

Full model name from the model list

Special note: If you select the OpenAI provider, only gpt series models are supported. For claude / gemini / deepseek / kimi / qwen / glm etc., select the corresponding model provider, set the request address (domain + /v1) and Key, then save. Switch to it on the session page.

Dify

Local Docker deployment · OpenAI-API-compatible model vendor configuration

① Install Docker: Download and install Docker Desktop. On first launch, skip login by clicking "Continue without signing in". For users in China, configure a domestic mirror source. Move the disk image location (Settings → Resources → Advanced) off the C drive to avoid filling it up.

② Get Dify Project: Clone via git or download ZIP and extract. For faster downloads in China, use: https://ghfast.top/https://github.com/langgenius/dify/archive/refs/heads/main.zip

③ Configure and Start: Navigate to the project root directory, copy .env.example to .env. Go to the dify/docker directory and start all services:

docker compose up -d

④ Access and Configure Models: Open http://localhost/ in your browser to create an administrator account and log in. Click the account avatar → Settings → Model vendor. Select and install "OpenAI-API-compatible", click to add a model, fill in the API Key, API URL (domain + /v1), and save.

API URL

https://api.we64.com/v1

API Key

The key we send you on WhatsApp after activation

Model

Full model name from the model list

Please Read: Data & Hosting Location

Important disclosure about where your API requests are processed.

Every model on we64.com runs on infrastructure and data centers located in China. Requests and the content you send to the API pass through servers there. We don't control or audit that infrastructure ourselves — we provide access to it.

This is fine for most everyday use — coding help, drafting, general chat, experimentation. But if you're working with highly confidential, regulated, or sensitive data (client data, health records, trade secrets, government or legal work), we'd honestly recommend using an official provider with data-residency guarantees in your own region instead.

Good for

• Coding help + debugging
• Drafting + general chat
• Experimentation + prototyping
• Non-sensitive data processing
• Personal + internal team tools

Don't use for

• Client / customer confidential data
• Health records (HIPAA / similar)
• Trade secrets + proprietary IP
• Government / legal / regulated work
• Anything with data-residency requirements

We're upfront about this: all infrastructure is hosted in China, and we don't control or audit it ourselves. For everyday tasks that's rarely an issue, but for confidential, regulated, or sensitive work, please consider an official provider with data-residency guarantees in your region instead.

Supported API features (all models)

Chat Completions

/v1/chat/completions

Streaming (SSE)

stream: true

JSON mode

response_format: json_object

Function calling

tools + tool_choice

Tool-use

Native for all models

Multi-turn

Full conversation memory

Max tokens

Up to model context limit

Need help integrating?

Email us for technical support — we reply within hours during business hours. Or chat with us on WhatsApp for instant help.

Email for support Chat on WhatsApp

Quick View

Technical Documentation

Base URL & Authentication

MCP Endpoint

Plug we64.com into your tools

Please Read: Data & Hosting Location

Supported API features (all models)

Need help integrating?