---
title: "Usage data is now queryable in the Alchemy CLI"
description: "Pull your team's compute and cost data from the terminal with alchemy usage summary and alchemy usage timeseries. Same data is available to AI agents through the Alchemy MCP Server."
---

# Usage data is now queryable in the Alchemy CLI

<ImageBlock
  src="https://media.alchemy.com/usage-data.png"
  alt="Usage Data in the Alchemy CLI"
  width={1920}
  height={900}
  priority
/>

Usage data is now queryable in the [Alchemy CLI](https://www.alchemy.com/docs/alchemy-cli#usage). You can pull your team's compute and cost data, both totals and a breakdown over time, straight from the terminal. The same data is available to AI agents through the [Alchemy MCP Server](https://www.alchemy.com/docs/alchemy-mcp-server), so an agent can see what your team is consuming without a custom integration.

Until now, checking usage meant logging into the [Dashboard](https://dashboard.alchemy.com/) and reading the charts manually. There was no first-class programmatic way to get at the numbers. Now there is.

## What you can do

Two commands cover most of what you'll reach for.

- `alchemy usage summary` gives you month-to-date, last 7 days, or last 30 days totals, in both compute units and USD.
- `alchemy usage timeseries` gives you the same numbers plotted over time, so you can see when traffic and spend moved.

Both render as a readable chart in your terminal. Add `--json` and you get structured output instead, ready to pipe into a script, a dashboard, or an alert.

<CodeSnippet
  language="bash"
  code={`alchemy usage summary
alchemy usage timeseries`}
/>

For machine-readable output:

<CodeSnippet
  language="bash"
  code={`alchemy usage summary --json
alchemy usage timeseries --json`}
/>

## Slice it the way the Dashboard does

You can group usage by network, app, method, or request type, and filter on any of them, so you can answer questions like which app is driving compute or which method spiked overnight. Pick hourly or daily granularity with `--granularity`. This is full parity with the Dashboard's usage view, which stays the visual source of truth.

A few examples:

<CodeSnippet
  language="bash"
  code={`alchemy usage summary --metrics amount,usd
alchemy usage timeseries --group-by network --granularity day
alchemy usage summary --app-ids <id> --methods eth_getLogs`}
/>

## Give your agents usage awareness

The same usage data is exposed through the Alchemy MCP Server as two tools: `get_usage_summary` and `get_usage_time_series`. Point an agent at your MCP server and it can reason about consumption and cost as part of its context, so an agent monitoring your infrastructure also knows what that infrastructure is costing.

If you already connected the MCP server to Cursor, Claude Code, or Codex — see the [MCP setup guide](https://www.alchemy.com/docs/alchemy-mcp-server) or our [MCP announcement](/blog/alchemy-mcp-server) — these tools show up alongside the rest of the Alchemy surface. Pair them with [Agent Skills](https://www.alchemy.com/docs/alchemy-agent-skills) if you want your agent to know how to call Alchemy APIs in code; MCP is what gives it live usage data in the conversation.

## Who this is for

If you're building agents or you just want programmatic, scriptable visibility into what your team consumes, this is for you. Monitor spend while you develop. Wire usage into an agent's context. Build your own cost dashboard or a budget alert. Track down which network, app, or method is responsible for a jump in compute. All of it now lives where you already work.

## Getting started

Make sure you're on Node 22 or later, then install or update the CLI:

<CodeSnippet
  language="bash"
  code={`npm install -g @alchemy/cli`}
/>

Then run `alchemy usage summary` and you're looking at your numbers. See the [usage docs](https://www.alchemy.com/docs/alchemy-cli#usage) for flags and filters, or our [12 things you can do faster with the Alchemy CLI](/blog/12-things-faster-with-alchemy-cli) walkthrough for the full command surface.

Usage in the CLI and the MCP server is one more step toward a platform that agents and developers can both operate directly. Update the CLI, run a command, and tell us what you'd want to see next.