> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lasso.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# How Lasso learns your providers

> Lasso watches every upstream answer and stops sending requests to providers that can't serve them, without you configuring their limits.

# How Lasso learns your providers

Every provider has limits nobody wrote down: methods it doesn't serve, how far
back it keeps state, how wide an `eth_getLogs` range it accepts. Lasso learns
these from the answers it already gets, then routes around them. You don't
configure them, and you don't change your requests.

## What it learns

For each provider, Lasso learns:

* **Methods it refuses.** A provider that keeps answering "method not supported"
  stops getting that method first.
* **How far back it serves.** A provider that only keeps recent state stops
  getting reads for older blocks first.
* **Its `eth_getLogs` range limit.** A provider that rejects wide ranges stops
  getting wide ranges first.

This runs on every profile: Lasso's managed pools and your Custom profiles.

## How it acts

Learning only changes the **order** in which Lasso tries providers. A provider
that can't serve a request is tried after the ones that can. It is never
removed, and Lasso always keeps the last route, so a request still has
somewhere to go even if everything learned turns out to be wrong.

## How it stays correct

* **It needs repeated evidence.** One odd answer doesn't create a limit. A range
  limit, for example, needs failures from two separate requests close together.
* **A success clears it.** If a provider later serves what it once refused, the
  learned limit goes away.
* **It fades.** Learned limits expire unless new failures keep renewing them, so
  a provider that upgrades its plan gets traffic back on its own.
* **It's scoped.** Knowledge belongs to one provider account and one method. One
  customer's credentials don't teach Lasso about another's.
* **It survives restarts.** Lasso restores what it learned when a node restarts.

## What you control on a Custom profile

* **See it.** Each provider's learned knowledge, with its source, is in the
  profile; see [Custom profiles](https://docs.lasso.sh/cloud/custom-profiles.md#lasso-learns-your-providers).
* **Declare what you know.** A limit you declare, such as your plan's log range,
  always outranks a learned one.
* **Correct it.** Reset a provider's knowledge, or pause one learned hint for an
  hour, from the API or the dashboard.

## Related

* [Error handling](https://docs.lasso.sh/cloud/reference/error-handling.md): what happens when a provider fails.
* [Indexer backfills](https://docs.lasso.sh/cloud/indexer-backfills.md): range limits in practice.


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