Your safe setup book for AI-coded projects.
Store API keys safely, generate clean env files, create MCP configs, and copy reliable prompts for Claude Code, Cursor, Hermes, and other coding agents.
Prompts use variable names and placeholders — not real secrets.
One safe setup book per project.
Secure project setup
Save API keys and service values in one workspace, encrypted in your browser and organized by environment.
Env builder
Select services and generate clean .env and .env.example files from the same source of truth.
MCP config helper
Create MCP snippets for Claude Code, Cursor, Hermes, VS Code, and Windsurf — without hand-building JSON.
Agent prompt packs
Copy setup and debug prompts that tell agents which variable names to use — never the real keys.
Docs-aware notes
Keep docs, auth links, and service caveats beside each service, and see when a page changes.
Safety guardrails
A sanitizer checks every generated output and blocks real secrets before you copy or download.
From new service to agent-ready setup in minutes.
- 01Create a projectName it after your app; local, preview, and production are set up for you.
- 02Add servicesOpenAI, Neon, Vercel, GitHub, Stripe, Resend, or your own.
- 03Fill required valuesStore real values safely — encrypted client-side, keyed to your vault password.
- 04Generate outputs.env, .env.example, MCP config, and setup prompts.
- 05Copy safelySend agents instructions with placeholders, not secrets.
Built to make unsafe copy/paste harder.
- Client-side encryption for stored secret values.
- .env.example output uses placeholders, not values.
- Agent prompts describe variable names, not raw keys.
- MCP snippets prefer env references and placeholders.
- A sanitizer checks generated outputs before copy or download.
EnvBook is practical safety infrastructure for builders — not a replacement for enterprise compliance work. If you lose your vault password, stored secrets cannot be recovered.