A2H: ask a human runs on your own machine — the client starts it, so there is no endpoint to ping. Last commit 8 Nov 2026.
Ask a human before acting: approvals by email, Telegram or push, signed answers, passkeys.
Nothing serious here today
Today is the operative word: we check A2H: ask a human every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Alternatives to A2H: ask a human
same job, measured the same wayHuman approval gate for AI agents. Critical actions approved or rejected by a human via Telegram.
Gate AI agent actions behind a consent policy with human approval via Telegram. Fail-closed.
Check an email before an AI agent sends it: send / ask a human / block. Detects prompt injection.
Check a vendor (name, domain, email) for fraud signals and sanctions before an agent pays it.
Ask a real, manually approved developer for help. Nothing is sent until you confirm by email.
Let AI agents notify you on Slack, Telegram, email, SMS, or push via an Alphorn webhook.
Human-in-the-loop for AI coding agents — ask questions, get approvals via Slack.
Crypto wallet: balances, P&L, history. Swaps, transfers and withdrawals need Telegram approval.
A2H: ask a human — questions
Answers built from our own checks of this server.