Keep Claude Code’s API keys out of .env files and chats.

Anthropic’s coding agent, in the terminal and in Claude Code cloud sessions. With Vando, your keys stay on your phone. When Claude Code needs one, your phone shows who is asking and why, and you give it just what it asked for, for one run or a set time, or you deny it with one tap.

The Claude Code guide, step by step

Checked against Claude Code’s own documentation on 2026-10-30; the guide links its sources.

Set up Claude Code in four steps.

  1. 1 · Install vando and connect it to your phoneIn your own terminal, not through Claude Code. vando connect shows a code; open Vando on your phone, tap Connect an AI, enter it and tap Connect. The code works once, for 10 minutes.
    $ npm i -g @getvando/cli --registry=https://registry.npmjs.org --@getvando:registry=https://registry.npmjs.org
    $ vando connect
  2. 2 · Tell it how to askvando setup writes the note into AGENTS.md and adds @AGENTS.md to CLAUDE.md, which Claude Code reads. It also writes Claude Code’s permission rules in .claude/settings.json: vando request, status and grants run without asking, vando run, vando get and vando env always ask. Other settings are left alone.
    vando setup
  3. 3 · Let vando run reach your phoneClaude Code asks before each command. Choose Yes for vando run, never “don’t ask again”: that would let any command run through vando run -- … skip Claude Code’s own check. Your phone asks as well, every time.
  4. 4 · The MCP tools, if you want themOptional, for the rare case Claude must read a value itself (a value a tool returns stays in the conversation). Start a new Claude Code session after adding it.
    claude mcp add vando -- vando mcp

Use Vando in every project.

Claude Code reads ~/.claude/CLAUDE.md in every project. Claude Code reads it at the start of every session, before the project’s CLAUDE.md; the files add up rather than override each other. Rules can also go in ~/.claude/rules/. (Claude Code’s memory docs.)

Put this note there, so Claude Code gets secrets through Vando everywhere, suggests vando add in a new project, and suggests moving a project’s .env files into Vando when it finds them:

## Secrets (Vando), in every project
- Never open, print, copy or edit `.env*` files or secret values, and never put a secret in chat.
- Run anything that needs a secret as `vando run --secret NAME -- <command>` (repeat `--secret` for more than one).
- In a new project, keep secrets in Vando from the start: suggest `vando add NAME` for each one, then `vando run` to use it.
- In a project that has `.env` files with values in them, suggest `vando init`: it moves them into Vando with one Save on the owner's phone, so the values stop living in plain files that you and other tools can read.
- Exit 75 means it is waiting for the owner's phone: say "Approve NAME on your phone, then I'll run it again", do other work, then run the same command again. After two waits, stop and tell the owner to open Vando on their phone (the request is in the Inbox).
- Exit 77 means the owner said no or the access ended: don't retry; ask the owner.

An example: A deploy from a cloud session, with a key for an hour.

You
Ask Claude Code, in a cloud session, to deploy the branch.
Claude Code runs
vando run --secret DEPLOY_TOKEN -- pnpm deploy
Your phone shows
“Claude Code · fix-login-flow · 4f2a wants to use DEPLOY_TOKEN”, with Once, 1 hour, 8 hours, 1 day. You tap Give for 1 hour, then your fingerprint.
Then
Every vando run in that session that asks for the same key gets it until the hour ends, without ringing your phone. The value is in the command’s environment, never in the chat.

In the cloud, or on another computer.

A Claude Code cloud session connects the same way: install vando there, run vando connect, enter the code on your phone. The session’s network blocks Vando’s domains until you allow them once in the environment’s network settings; vando names the domain when it is blocked. The connection’s key lives only in that session, and secrets still ask your phone.

Good to know with Claude Code.

Questions.

Can Claude Code use my keys in a cloud session?

Yes. A cloud session starts with no access to Vando’s server, so you allow its domain once in that environment’s network settings (vando names it when it is blocked). After that the keys stay in your vault and your phone asks each time.

Where does Claude Code find out how to ask?

From the note vando setup writes in AGENTS.md, which Claude Code reads through the @AGENTS.md line in CLAUDE.md. For every project at once, put the every-project note in ~/.claude/CLAUDE.md.

What stops Claude Code from reading the key anyway?

With vando run, the key goes into that one command’s environment and anything the program prints of it is shown to Claude as [vando:NAME]. Only the optional get_secret tool puts a value in the conversation, and only after you tap Give.

What Vando is, in short.

One app on your phone for your passwords, 2FA codes, recovery kits, files and developer secrets. Each thing gets its own lock: Easy, Just me, Two devices or Fort Knox, from a quick tap up to your fingerprint plus a security key. Your laptop and your AI tools hold no keys; when one needs a login or a key, your phone shows who is asking and for what, and you say yes or no there. Every yes is signed on your phone and kept on record.

How the locks work → · The detail for developers → · Every AI tool you can set up with Vando → · Get the app →