About XPEACH and 🍑Code site cluster
Welcome to the XPEACH Unified Pass Center (https://sso.xpeach.codes).
XPEACH is a personal independent technical service cluster for developers and AI engineering scenarios. It is composed of XPEACH SSO unified identity authentication center, 🍑Code multi-line high-speed model gateway and XPeach Harness intelligent programming workbench. It provides a development experience of "one pass for the entire cluster and one set of quotas for seamless cross-end calls".
1. Core product and service matrix
1. XPEACH SSO · Unified Pass Hub (sso.xpeach.codes)
As the only identity authentication and security base for the entire cluster, XPEACH SSO is responsible for managing your core account assets and cross-site authorization:
- Global immutable UUID: Each pass is assigned a unique 16-bit underlying identifier in the format of
****-****-****-****, which serves as the unique identity primary key of each subsite and client in the entire cluster. - Modern passwordless and multi-factor authentication: native support for Passkey (FIDO2 / WebAuthn pass key) fingerprint/face anti-phishing login, TOTP dynamic password and one-time recovery code.
- Multi-channel federated identity binding: Supports one-click quick login by binding external developer accounts such as GitHub, Google, LinuxDO, Discord, etc.
- Single sign-on and device flow bridging: Provides web page single sign-on based on the standard OpenID Connect (OIDC)/OAuth 2.0 protocol (PKCE is forced to be enabled), and implements one-click credential automatic bridging between the desktop and CLI tools through 8-digit device codes (Device Flow).
- Cross-site Taozi (🍑) ledger and invitation system: unified accounting recharge, CDK exchange, cross-site quota synchronization and cluster-wide universal invitation code (
aff_code).
2. 🍑Code · High-speed AI model aggregation gateway (xpeach.codes)
Model routing services for AI programming, agent orchestration and high-frequency inference scenario optimization:
- Full cutting-edge model matrix: Aggregates mainstream cutting-edge codes and inference models such as Claude, GPT, Gemini, DeepSeek, Kimi, Grok, etc., and natively supports dual-protocol offloading of
openai-completionsandanthropic-messages. - Three-line intelligent acceleration: Provides main station (https://xpeach.codes), high-speed (https://fast.xpeach.codes) and express (https://ultrafast.xpeach.codes) three transfer acceleration lines, supporting grouping (such as
SWE,GPT|Standard,Claude|Standard) can be switched flexibly. - Transparent metering and real-time risk control: Provide detailed query of token consumption, delay time and grouping rate on a transaction-by-transaction basis.
3. XPeach Harness (xph) · Fully plug-in AI programming agent workbench
The cross-platform agent development workbench (Agentic Coding Workbench) specially built for the 🍑Code ecosystem provides macOS/Windows/Linux native desktop client, Web workbench, Headless CLI and SDK:
- SSO one-click authorization and zero configuration to get started: Click "Use XPeach SSO one-click login" to automatically complete 🍑Code account instant opening (JIT Provisioning), system access token issuance and exclusive
sk-key binding; built-in three-line continuous detection and automatic preferred switching. - Full-stack agent engineering tool chain: built-in controlled file reading and writing and patch editing (
fs), persistent terminal command execution (shell), LSP code diagnosis (lsp), web search and crawling (web), browser automation (browser-use), desktop operation (computer-use), sub-agent concurrent collaboration (subagent) and programmatic tool invocation (PTC Runtime). - Zero-threshold ecological migration: Supports one-click import of Claude Code, Codex, Gemini CLI, Antigravity, ChatGPT and DeepSeek Harness's historical sessions, Skills skill packs, Rules global rules and MCP server configuration that have been installed on the machine in a local read-only, one-way manner.
- Bill-level real-time usage capsule: The 🍑 status capsule in the corner of the workbench displays the remaining balance of the account, today's consumption and the trend in the past 7 days in real time.
2. Security and Privacy Design Principles
- High-Strength Encrypted Storage: Login passwords are stored using irreversible salted hashes; TOTP keys, third-party binding tokens, and subsite synchronization keys are encrypted using
AES-256-GCM; API Keys and session tokens only storeSHA-256digests. - Zero biometric upload: Passkey fingerprint and face verification are completely completed within the security hardware chip (Secure Enclave/TPM) of your local device, and the platform never touches the original biometric data.
- Local workspace isolation: XPeach Harness stores all local sessions and credentials independently in the local
~/.xphdirectory (credential files are strictly limited to0600permissions); except for the model inference context that you actively initiate, local unrelated code repositories will not be silently uploaded in the background. - Millisecond-level dynamic revocation: When SSO changes passwords, logs out of devices, or triggers security bans, the full cluster subsite session and client local credentials are revoked through the signature Webhook and real-time verification mechanism.
- Not used for model training: This platform will never actively use your prompt words, upload code or model response content to train its own general basic large model.
3. Quick navigation and contact information
- Unified Pass Center: https://sso.xpeach.codes
- 🍑Code main site and console: https://xpeach.codes
- Legal and Compliance Documents: User Service Agreement · Privacy Policy
- Official contact email: xpeachcodes@gmail.com
XPEACH SSO