# API Rate Limiting Many input connectors let you cap the rate at which Monad makes outbound requests to the source API. Setting a limit keeps a connector comfortably under the provider's quota, avoiding `429 Too Many Requests` errors and the throttling/backoff delays they cause. ## How it works When you set a rate limit, Monad applies a token-bucket limiter to the connector's outbound HTTP requests, smoothing them to the rate you choose. This is _proactive_: it prevents you from hitting the provider's limit in the first place. Independently, Monad always honors a provider's `Retry-After` header and backs off on a `429`, so a connector stays correct even without a configured limit. ## Configuring a limit The **API Rate Limit** setting is an object with two fields: | Field | Description | | ----- | ------------------------------------------------------------------- | | Rate | A positive number — how many requests are allowed per unit of time. | | Unit | `per_second` (default), `per_minute`, or `per_hour`. | For example, _Rate = 30, Unit = per_minute_ limits the connector to 30 requests/minute (0.5/sec). ## Defaults and limits - **Leave it blank** to use the connector's default behavior. For most connectors this means unthrottled (Monad still backs off on `429`); some connectors apply a built-in default — see the connector's own page. - Each connector enforces a **maximum** based on the provider's documented API ceiling. A value above that maximum is rejected when you save the configuration. - The platform **minimum** is 0.1 requests/second; a limit slower than that is rejected. ## Choosing a value - Start from the provider's documented rate limit (linked on each connector's page). - If your API token is shared with other tools, set a limit below the provider's ceiling to leave headroom. - Plan tiers matter: many providers' limits depend on your subscription, so the connector's maximum is a conservative ceiling, not a guarantee for your account.