What is Zro?
Zro is a private inference endpoint built specifically for coding agents. It serves open-weight coding models through a fast, multi-region API with zero request retention and no training on customer data.
Key Features
- Open-weight coding models: Access GLM-5.3 Flash, DeepSeek V4.1 Flash, and Kimi K3 through a single endpoint.
- Privacy-first design: Prompt and completion bodies are not retained after inference, and customer data is never used for training, fine-tuning, evaluations, analytics, or dataset creation.
- Multi-region infrastructure: Inference runs across multiple regions on privacy-forward infrastructure.
- Optimized for long-context coding: Tuned for long-context, multi-turn coding sessions using HyperQuant compression, custom attention kernels, and hardware-aware deployment across AMD, NVIDIA, and TPU.
- Fast streaming performance: Built for responsive streaming inference so agents and production apps don't trade speed for privacy.
Getting Started
Install the npm package, log in once, and launch supported coding tools with temporary session config:
npm install -g @moonmath-ai/zro
zro login
zro launch claude
zro launch codexYou can also specify a model, for example zro launch claude --model glm-5.3 or zro launch codex --model kimi-k3.
Integrations
Zro works with Claude Code, Codex CLI, Cursor, Cline, Kilo Code, Oh My Pi, Pi, OpenClaw, OpenCode, Grok Build, Hermes, and Prime Agent. Manual setup guidance is available for Cursor and Cline.
Compatibility
Zro exposes OpenAI-compatible access for chat completions, so existing clients and agent tools can point at the Zro base URL. It also supports Anthropic-compatible Messages requests at /v1/messages for tools that expect that API shape.
Billing
Plans start at $20/month for $60 of inference spend. Usage packs are available without a subscription and expire after 90 days, while plan spend resets monthly.