# Agent Vault

> Give an AI agent a card and an account without giving it your credentials.

- [What is Agent Vault](https://docs.agentvault.io/index.md): Let an AI agent shop and sign in for you without ever giving it your card number or your passwords.
- [Quickstart](https://docs.agentvault.io/get-started/quickstart.md): Go from sign-up to an agent that completes its first checkout, in about 15 minutes.
- [Core concepts](https://docs.agentvault.io/get-started/concepts.md): The nine terms the rest of the documentation builds on, each in a few sentences.
- [How it works](https://docs.agentvault.io/get-started/how-it-works.md): Follow a checkout from your agent to the merchant, and see exactly where the real card number appears and what is checked first.
- [Choose an integration](https://docs.agentvault.io/get-started/choose-an-integration.md): Pick between the MCP server, the Playwright SDK and the agent skill in under a minute.
- [FAQ](https://docs.agentvault.io/get-started/faq.md): Short answers to the questions new account holders and developers ask first.
- [Run a checkout end to end](https://docs.agentvault.io/guides/first-checkout.md): Take an agent from a product page to a placed order, with the account holder's limits applied and every human step in the right place.
- [Sign in to a site](https://docs.agentvault.io/guides/sign-in.md): Sign an agent in with a stored login, including email-first pages and authenticator codes, without the agent ever seeing the password.
- [Let an agent open an account](https://docs.agentvault.io/guides/agent-signup.md): Have an agent sign the account holder up for a site, with their approval, using a password the vault generates that neither the account holder nor the agent receives.
- [Handle challenges](https://docs.agentvault.io/guides/challenges.md): Detect a 3-D Secure, one-time code or CAPTCHA step, or a site that silently blocks the agent, and hand the live page to the account holder to complete.
- [Live CVV entry](https://docs.agentvault.io/guides/live-cvv.md): Fill a card's security code without it ever being stored, and handle the cases where the account holder cannot supply it in time.
- [Page hints](https://docs.agentvault.io/guides/page-hints.md): Point Agent Vault at a checkout total it cannot find, so the account holder's limits can be checked, and store an authenticator secret the agent never sees.
- [Confirming what happened](https://docs.agentvault.io/guides/confirming-outcomes.md): Check whether an order actually landed, using the live page, the session's run report and the merchant's receipt, before you tell anyone their card was charged.
- [Dashboard overview](https://docs.agentvault.io/dashboard/overview.md): Finish setup, check this week's spend, and stop every agent at once from the Overview page.
- [Cards](https://docs.agentvault.io/dashboard/cards.md): Add the payment cards your agents pay with, without them ever seeing the number.
- [Agent cards](https://docs.agentvault.io/dashboard/agent-cards.md): Set how much each agent can spend, where it can pay, and which purchases need your approval.
- [Logins](https://docs.agentvault.io/dashboard/logins.md): Save site sign-ins so your agents can sign in without ever seeing your password.
- [Identity and addresses](https://docs.agentvault.io/dashboard/identity-and-addresses.md): Save the name, contact details and addresses your agent fills into checkouts and sign-up forms.
- [Destinations](https://docs.agentvault.io/dashboard/destinations.md): Choose the sites where your agent may use your saved identity, addresses and card details.
- [API keys](https://docs.agentvault.io/dashboard/api-keys.md): Create, limit and revoke the keys your agents use to connect to Agent Vault, and the apps connected through the hosted MCP server.
- [Approvals and notifications](https://docs.agentvault.io/dashboard/approvals-and-notifications.md): Answer the requests that pause your agent, and choose how Agent Vault reaches you.
- [Activity](https://docs.agentvault.io/dashboard/activity.md): See what your agents filled, where, what was refused and why, from a record your agents cannot edit.
- [Billing and plans](https://docs.agentvault.io/dashboard/billing.md): Compare plans, see your usage, upgrade, and know what happens when you reach a plan limit.
- [Data and deletion](https://docs.agentvault.io/dashboard/data-and-deletion.md): See what Agent Vault keeps about you and for how long, export a copy, or delete your account.
- [Connect your agent](https://docs.agentvault.io/agents/overview.md): Connect Meta Muse, Grok Bot, OpenClaw, Claude Code, Hermes, Claude Desktop, Antigravity or Codex to Agent Vault in about 2 minutes.
- [Connect Meta Muse](https://docs.agentvault.io/agents/meta-muse.md): Add Agent Vault to Meta Muse as a custom connector and sign in, in about 2 minutes.
- [Connect Grok Bot](https://docs.agentvault.io/agents/grok-bot.md): Add Agent Vault to Grok Bot as a custom MCP server and authorize it, in about 2 minutes.
- [Connect OpenClaw](https://docs.agentvault.io/agents/openclaw.md): Add Agent Vault to OpenClaw with one command and sign in, in about 2 minutes.
- [Connect Claude Code](https://docs.agentvault.io/agents/claude-code.md): Add Agent Vault to Claude Code with one command and sign in with /mcp, in about 2 minutes.
- [Connect Hermes](https://docs.agentvault.io/agents/hermes.md): Add Agent Vault to Hermes with one command and sign in, in about 2 minutes.
- [Connect Claude Desktop](https://docs.agentvault.io/agents/claude-desktop.md): Add Agent Vault to Claude Desktop or claude.ai as a custom connector and sign in, in about 3 minutes.
- [Connect Antigravity](https://docs.agentvault.io/agents/antigravity.md): Add Agent Vault to Google Antigravity and authenticate it in Customizations, in about 3 minutes.
- [Connect Codex](https://docs.agentvault.io/agents/codex.md): Add Agent Vault to OpenAI Codex with one command that also starts sign-in, in about 2 minutes.
- [Set up the MCP server](https://docs.agentvault.io/mcp/setup.md): Add the Agent Vault MCP server to Claude Code, Claude Desktop, Cursor or Codex: sign in to the hosted server, or add one local config block.
- [Tool reference](https://docs.agentvault.io/mcp/tool-reference.md): Every Agent Vault MCP tool, with its parameters, what it returns and when it refuses.
- [Rules of engagement](https://docs.agentvault.io/mcp/rules-of-engagement.md): The nine rules an agent using Agent Vault follows, with the reason behind each.
- [Install the SDK](https://docs.agentvault.io/sdk/install.md): Install @agentvault/playwright, configure it, and run your first proxied checkout.
- [SDK API reference](https://docs.agentvault.io/sdk/api-reference.md): Every export of @agentvault/playwright: the catalog, signup and registry helpers, and the proxied browser object model.
- [Sessions and timeouts](https://docs.agentvault.io/sdk/sessions-and-timeouts.md): How long a proxy session lives, which errors end it, and how to recover from each.
- [Egress and tunnelling](https://docs.agentvault.io/sdk/tunnel.md): Route the proxy browser's traffic out through your own IP address, and choose what happens when the tunnel is unavailable.
- [Error codes](https://docs.agentvault.io/sdk/error-codes.md): Every error code the SDK and MCP server can return, what it means, and whether retrying is safe.
- [Checkout recipe registry](https://docs.agentvault.io/registry/overview.md): An open registry that records how real merchant checkouts work, so an agent can follow a known path instead of improvising.
- [Recipe format](https://docs.agentvault.io/registry/recipe-format.md): Every field of a platform, merchant and signup recipe, with the rules a valid recipe must follow.
- [Recipe statuses](https://docs.agentvault.io/registry/statuses.md): What each recipe status claims, the proof it needs, and who can set it.
- [Look up a recipe](https://docs.agentvault.io/registry/lookup.md): Ask the registry what is already known about a site before your agent touches the cart.
- [Contribute a recipe](https://docs.agentvault.io/registry/contribute.md): Submit a recipe through the API, read the screening verdict, and fix a rejection.
- [Checkout recipes skill](https://docs.agentvault.io/registry/skill.md): Install the registry as an agent skill so your agent looks a site up, follows the recipe and contributes what it learns.
- [Security model](https://docs.agentvault.io/security/model.md): The threat model behind Agent Vault: what it protects, who it defends against, and where its protection ends.
- [What the agent can and cannot see](https://docs.agentvault.io/security/agent-context.md): The exact read surface available to an agent before and after its first token fill, and how to verify it.
- [Card storage](https://docs.agentvault.io/security/card-storage.md): Where a card number is stored, how it reaches a merchant form, and why it never reaches Agent Vault's backend.
- [Attested detokenization](https://docs.agentvault.io/security/attested-detokenization.md): Revealing a card number only inside a hardware-attested enclave: current status, design, limitations and error codes.
- [CVV policy](https://docs.agentvault.io/security/cvv-policy.md): Why the card security code is never stored, how a checkout that needs one is completed, and how to verify it.
- [URL locking](https://docs.agentvault.io/security/url-locking.md): How each kind of token is limited to the sites it may be used on, why logins fail closed and cards follow the agent card rules, and how to verify it.
- [TOTP handling](https://docs.agentvault.io/security/totp.md): How an agent completes an authenticator-app step without a stored secret ever leaving the backend, how agent enrolment hands a new secret to it, and how to verify it.
- [Screenshot redaction](https://docs.agentvault.io/security/redaction.md): Why every screenshot is black after a token fill, why most sessions store no receipt image, and what the audit trail never contains.
- [Reporting a vulnerability](https://docs.agentvault.io/security/disclosure.md): How to report a security issue in Agent Vault, what is in scope, how to test responsibly and what to expect after you report.
- [HTTP API](https://docs.agentvault.io/reference/http-api.md): Every Agent Vault endpoint you can call without the SDK or the MCP server, with parameters, responses and errors.
- [Token catalog](https://docs.agentvault.io/reference/token-catalog.md): The exact shape of the token catalog, field by field, including what each optional field's absence means.
- [Token grammar](https://docs.agentvault.io/reference/token-grammar.md): Every mock token shape, what it stands for, and where it resolves.
- [Environment variables](https://docs.agentvault.io/reference/environment-variables.md): Every environment variable the SDK and the MCP server read, with defaults.
- [Webhooks](https://docs.agentvault.io/reference/webhooks.md): Receive an agent card's events on your own HTTPS endpoint, and verify each delivery came from Agent Vault.
- [Limits and timeouts](https://docs.agentvault.io/reference/limits.md): Every plan quota, rate limit, ceiling and deadline in one place, and whether waiting clears it.
- [Troubleshooting](https://docs.agentvault.io/reference/troubleshooting.md): Find the symptom your agent hit, see what causes it, and apply the fix.
- [Glossary](https://docs.agentvault.io/reference/glossary.md): Every term the Agent Vault documentation uses, in plain language first and then precisely.

## Optional

- [Dashboard](https://app.agentvault.io/dashboard)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.