All articles

How ShareMyTokens works

Follow a share link from setup to first use, and learn what the host, guest, and AI provider each control.

By ShareMyTokens ·

One link, three roles

ShareMyTokens lets a host grant temporary access to an AI model using a private URL. The host connects a provider API key in a vault and decides how many tokens a link can use. The guest opens the link and uses the access the host assigned. The AI provider supplies the model and bills the account associated with the provider key.

The distinction matters when you invite someone to try a model. You can give them a working chat with a specific budget and expiration without copying a provider credential into an email or asking them to configure a developer account. The provider key stays on the service side. Each link gives the host a separate place to track usage.

Create and test your first link

Sign in, open Vaults, and add a key from a provider supported by your plan. In Share Links, choose the provider and set a token budget and expiry that fit the task. For a first test, pick a small budget and a short expiry. Share the generated URL through a channel appropriate for the people who should have access.

Choose an open link when anyone who has the URL can participate. Choose a claimable link when access should go to the first signed-in recipient who claims it. A claimable link still needs careful delivery because the first eligible person to open and claim it gets the access. Where your plan supports it, you can add a password at creation.

Check the guest flow before distributing the link. Use a separate browser profile for an open link, or create a separate test link if you need to exercise the claim flow. Make one short request and check the usage in the host dashboard. Revoke the test link afterward, then create the link you intend to distribute.

When the recipient wants an API

Eligible Pro claimable links can provide API access when the host enables it. The signed-in claimant can generate a ShareMyTokens key for compatible clients. The chat and API use the same link budget, so switching clients does not create another allocation. Check the developer documentation for supported providers and request formats before planning an integration.

Further reading