Keep Aider’s API keys out of .env files and chats.
The open-source terminal pair programmer. With Vando, your keys stay on your phone. When Aider 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.
Checked against Aider’s own documentation on 2026-11-10; the guide links its sources.
Set up Aider in four steps.
-
1 · Install vando and connect it to your phoneIn your own terminal, not through Aider.
vando connectshows 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 · Tell it how to askAider loads no rules file by itself. Put the note in
AGENTS.md(orCONVENTIONS.md) and load it read-only: addread: AGENTS.mdto.aider.conf.ymlin the repository root, or start withaider --read AGENTS.md.read: AGENTS.md
-
3 · Let vando run reach your phoneAider has no sandbox and no allow list: its commands run as you, with your network, home folder and keychain. It asks
Run shell command?before each command it suggests; answer Y. You can also run one yourself with/run, and answer N to “Add command output to the chat?” for the connect code, so it isn’t sent to the model. With--yes-always, Aider answers “no” to suggested shell commands, so an unattended Aider never runsvando run; that is a limit of Aider, not of Vando./run vando connect /run vando run --secret OPENAI_API_KEY -- pytest
-
4 · The MCP tools, if you want themAider has no MCP support, so Vando’s MCP tools (
get_secretand the rest) aren’t available with it. Everything goes throughvando run, which keeps secrets out of the chat anyway.
Use Vando in every project.
Aider has no user-level instructions file. The nearest thing is read: in a home-folder ~/.aider.conf.yml, which Aider loads before the repository’s and the current folder’s config files (files loaded last win). Point it at a file holding the every-project note, or keep the note in each project’s AGENTS.md.
Put this note there, so Aider 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: Running the test suite that needs an API key.
- You
- Ask Aider to run the tests, or type the command yourself.
- You or Aider run
vando run --secret OPENAI_API_KEY -- pytest- Aider shows
Run shell command?for a suggested command: answer Y. Then Vando prints “Waiting for your phone”.- Your phone shows
- “Aider · MacBook wants to use OPENAI_API_KEY”. Give once; pytest runs with the key in its environment.
In the cloud, or on another computer.
Aider has no hosted agent: it runs wherever you start it. Use Vando with Aider on a computer you sit at; in CI, --yes-always stops suggested commands.
Good to know with Aider.
- Aider passes a command’s output to the model, not its exit code, so the note tells the model to read what
vandoprints and run the same command again when it says it is waiting for the phone. - There is no automatic retry: you or the model must run the command again after exit 75.
- Keep Vando-managed secrets out of Aider’s own
.env: Aider loads it into its environment, and every command it runs inherits it. - No MCP means no
get_secret; that is fine,vando runis the road that keeps values out of the chat.
Questions.
Does Aider work with Vando’s MCP tools?
No. Aider has no MCP support, so get_secret and the others aren’t available. Use vando run, which gives the key to one command and never to the chat.
How does Aider know to run the command again after a wait?
Aider gives the model a command’s output but not its exit code, so the note says: read what vando prints, and when it says it is waiting for the phone, run the same command again. Nothing retries by itself.
Can I run Aider unattended with Vando?
Not for secrets. With --yes-always, Aider answers no to the shell commands it suggests, so vando run never runs. That is Aider’s rule, and it keeps an unattended Aider from spending your keys.
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 →