> ## 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.

# Serving access and pricing

> What serves each request (premium, public or Custom), what it costs in USD, and how compute units measure included usage

# Serving access and pricing

Every request Lasso serves has one **serving access**: `premium`, `public` or
`custom`. It decides which providers answer, what the request costs, which
strategies apply and which rate limit counts it, and every response tells you
which one served it. Read this page to choose a pool and predict your bill;
read [balance and payment fallback](/cloud/balance-and-fallback) for what
happens when money runs out.

```bash theme={null}
curl -sS -D - -o /dev/null "$LASSO_RPC_URL/fastest/base" \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"eth_blockNumber","params":[]}' \
  | grep -iE '^x-lasso-(access|usd|cu|fallback)'
```

```http theme={null}
x-lasso-access: premium
x-lasso-cu: 2
x-lasso-usd: 0.00001
```

## The three kinds of serving access

| Serving access | Providers | Strategies | Price | Rate limit |
| - | - | - | - | - |
| `premium` | Lasso's premium pool | All four | Charged per request | 100 RPS |
| `public` | Lasso's public pool | `load-balanced` only | Uncharged | 30 RPS, bursts to 60 |
| `custom` | Your Custom profile's providers | All four | Uncharged; your Custom subscription covers it | 1,000 RPS per profile |

Your URL names the profile, and your account's payment state decides how it is
served:

| The URL names | Served as |
| - | - |
| `premium`, or no profile with a `premium` default | `premium` while included usage or balance covers it; otherwise `public`, marked with `x-lasso-fallback` (see [balance and payment fallback](/cloud/balance-and-fallback)) |
| `public` | `public`, uncharged. It serves while the account has any spending power: balance, included usage, a payment grace window or a free allowance. An account with none gets 402 `balance_exhausted` on every profile, `public` included |
| A Custom profile | `custom` while your account has Custom access; otherwise `premium`, marked `x-lasso-fallback: profile` |

Choose `public` explicitly for traffic that doesn't need premium providers:
background jobs or development. Same key, different path
segment:

```ts theme={null}
const premium = `${process.env.LASSO_RPC_URL}/fastest/base`;
const background = `${process.env.LASSO_RPC_URL}/profile/public/base`;
```

## What a premium request costs

A request's size in **compute units (CU)** is its **method weight** times its
**strategy factor**. Its price is that CU at a published USD rate. A cheap read
like `eth_blockNumber` weighs less than `eth_getLogs`, and a trace weighs
more. The strategy factor is the trade you make between latency and cost:

| Strategy | Factor |
| - | - |
| `load-balanced` | 1x |
| `latency-weighted` | 1.5x |
| `fastest` | 2x |
| `priority` | 1x |

Every method's weight and price is in
[chains and prices](/cloud/reference/chains-and-prices), generated from the
live [`agent.json`](https://lasso.sh/agent.json). Every response states its
size in `x-lasso-cu` and its charge in `x-lasso-usd`, so nothing is a surprise.
A batch is priced per item, a WebSocket request costs the same as over HTTP,
and each subscription notification is priced as an `eth_subscription` entry.

Money is always in USD: prices, balances, purchases and charges. Included
usage, from free access or a subscription, is counted in CU per month.

Lasso charges only for a response a provider delivered, whether a result or a
JSON-RPC error. A failure inside Lasso with no provider response, and any
request refused before routing, is free.

### Current tariff

| Item | Value |
| - | - |
| Rate | \$0.000005 per CU |
| Strategy factors | `load-balanced` 1x, `latency-weighted` 1.5x, `fastest` 2x, `priority` 1x |
| Subscription notification | 3 CU |
| Starting grant for a new provisional account | \$0.025 |
| Free access | 1,000,000 CU per month |
| Pro included usage | 65,000,000 CU per month |

Subscription prices are on [lasso.sh/pricing](https://lasso.sh/pricing).
`agent.json` publishes the live values agents act on.

## Subscriptions and free access

| | What it gives your account | How you get it |
| - | - | - |
| **Free access** | A monthly CU allowance served on public providers, uncharged | A person on the account sponsors it; each person sponsors up to three accounts |
| **Pro** | A monthly CU allowance on premium providers; beyond it, requests draw on the balance | Subscribe in the dashboard |
| **Custom** | Routing over your own providers on up to five Custom profiles, unmetered | Subscribe in the dashboard |

Free access belongs to accounts that a person has joined. When you sign in
and create or claim your first account, you sponsor it, and you can sponsor
two more. A provisional account an agent created has no free access until a
person claims it.

Current subscription prices and allowances are on
[lasso.sh/pricing](https://lasso.sh/pricing) and in `agent.json`.

## Rate limits

Limits apply per account, not per key:

* One **managed bucket** holds every `premium` and `public` request from all
  of the account's keys: 100 RPS with bursts to 200 while the account has
  premium access, or 30 RPS with bursts to 60 while payment fallback or free
  access serves it.
* `public` requests also draw on a **public bucket** inside it: 30 RPS with
  bursts to 60. Public traffic can never take more than 30 of your 100
  premium RPS.
* Each Custom profile has its own 1,000 RPS bucket while it serves `custom`.

Every response carries the IETF `RateLimit-Policy` and `RateLimit` headers,
named after the bucket that bound the request: `managed`, `public` or the
Custom profile's slug. `"public";q=60;w=2` means 60 requests per two seconds. Over the
limit, Lasso answers 429 `rate_limit_exceeded` with `Retry-After`. Need more? Ask
Lasso support for an account override.

## Limits

* `public` serves `load-balanced` only. A request for another strategy is
  served as `load-balanced`, and `lasso_meta.strategy` reports what was used.
  The same holds for any request served by payment fallback or free access.
* Public providers' method coverage and historical depth can be narrower than
  the premium pool's.
* Custom traffic has no request price; your providers bill you as usual.

## Next

* [Balance and payment fallback](/cloud/balance-and-fallback): where the money
  comes from and what happens at zero.
* [Keys, URLs and profiles](/cloud/keys-and-profiles): default profiles and
  scoped keys.
* [Chains and prices](/cloud/reference/chains-and-prices): the live price
  table.


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