How to check RPC usage from the terminal with the Alchemy CLI

The Alchemy CLI lets you check RPC usage from the terminal. alchemy usage summary shows your team's compute units and USD spend in one command, and alchemy usage timeseries breaks that usage down by network, app, method, or request type. The same CLI manages your apps, allowlists, and webhooks.

Checking usage no longer means opening the Dashboard. Here's how the two usage commands work, how to find what's driving compute, what else the CLI manages, and how to script usage checks or hand them to a coding agent.

How do you check RPC usage from the terminal?

Sign in once, then ask for the summary.

alchemy auth opens a browser so you can sign in with your Alchemy account, and the usage commands reuse that login, so there's no API key to copy. On a remote machine, a CI runner, or an SSH session, run alchemy auth login --device-code and approve the short code from a browser on any device.

alchemy usage summary returns three totals for your team: month to date, the last 7 days, and the last 30 days. Each total comes in compute units and in USD, next to the start and end of your current billing period. The default output is human-readable. Add --json when a script or another tool needs the numbers.

Start with the summary for a quick read on the billing period. When a number is higher than you expected, the time series shows where it came from.

How do you see which app, network, or method is driving usage?

alchemy usage timeseries splits usage into daily or hourly buckets. It needs a start date, and it groups by up to three dimensions: app, network, method, and requestType.

Filters narrow the data before it's grouped. Each one takes comma-separated values, and they combine with AND.

--app-ids filters by app, and the request types are http, websocket, webhook, and grpc. The time series reports compute units only, so USD totals come from usage summary.

How far back the time series reaches depends on your plan, and one request covers at most 30 days.

If you ask for an earlier start date, the CLI clamps it to your plan's limit and shows both the requested and the returned range. When a total jumps, group by method first, then filter to the app and network that sent those calls.

What else can you manage from the Alchemy CLI?

The CLI covers the account and infrastructure work you'd otherwise click through in the Alchemy Dashboard.

• List apps and pick a default. alchemy app list and alchemy app select <id>
• Create an app. alchemy app create --name <name> --networks <ids>
• Restrict an app. alchemy app networks, alchemy app origin-allowlist, alchemy app ip-allowlist, and alchemy app address-allowlist
• Manage webhooks. alchemy webhook list and alchemy webhook create --body <json>
• Check your setup. alchemy auth status and alchemy doctor

Commands that change an app or a webhook accept --dry-run, so you can preview a change before it applies. The webhook commands also need a webhook API key from the Dashboard, which you set once with alchemy config set webhook-api-key <key>. Our Alchemy Webhooks page covers what each webhook type tracks.

The CLI reads chain data and sends transactions too. alchemy evm data balance vitalik.eth and alchemy evm block latest return results without a script, and Agent Wallets for the Alchemy CLI add signing and sending. For the rest of the command surface, see 12 things you can do faster with the Alchemy CLI.

Creating an app, restricting what it accepts, and checking the usage it generates can all happen in one terminal session.

How do you script usage checks or hand them to an agent?

Two global flags make any command safe to automate. --json prints structured output (piped output switches to JSON on its own), and --no-interactive stops the CLI from waiting on a prompt.

A budget check for cron or CI is one jq expression. This one exits with status 1 once month-to-date spend passes 500 USD.

Coding agents that run terminal commands, such as Claude Code, Cursor, and Codex, can call the same commands. Hand the agent the CLI's own instructions first.

The output lists the command tree, auth methods, error codes with retry guidance, and runnable examples, ready to drop into an agent's system prompt. We also ship two other agent surfaces. Agent Skills teach an agent to write Alchemy API calls in code, and the Alchemy MCP Server at https://mcp.alchemy.com/mcp gives it chain data and app management inside the conversation. Our MCP server announcement covers the setup for each client.

The MCP server exposes the same usage data as two tools, get_usage_summary and get_usage_time_series. An agent connected to it in Cursor, Claude Code, or Codex can check usage and cost mid-conversation, without running the CLI.

An agent that can read usage can flag a cost spike while it's still working on the code that caused it.

How do you install the Alchemy CLI?

Install it from npm, sign in, and run your first usage check.

If the CLI is already installed, the same command updates it. alchemy doctor reports anything missing from your setup and prints the command that fixes it. The usage section of the CLI docs lists every flag and filter.

Frequently asked questions

Is there a CLI for managing blockchain infrastructure and checking RPC usage?

Yes. The Alchemy CLI manages Alchemy apps, allowlists, and webhooks from the terminal, and alchemy usage summary shows month-to-date, 7-day, and 30-day usage in compute units and USD. alchemy usage timeseries breaks usage down by network, app, method, or request type. Install it with npm i -g @alchemy/cli@latest.

How do I check my Alchemy compute unit usage from the command line?

Install the Alchemy CLI, sign in with alchemy auth, and run alchemy usage summary. It returns month-to-date, last-7-day, and last-30-day totals in compute units and USD for your team. Add --json for structured output you can pipe into jq, a dashboard, or an alert.

Can I see RPC usage by method, network, or app?

Yes. alchemy usage timeseries --start-date <date> --group-by method groups usage by method, and you can group by up to three of app, network, method, and requestType at once. Filters such as --networks, --methods, --app-ids, and --request-types narrow the data first. Pick hourly or daily buckets with --granularity.

How far back can I query usage from the Alchemy CLI?

The usage time series looks back 1 day on the Free plan, 7 days on Pay as You Go, and up to 90 days on Enterprise when it's enabled for your team. Each request covers at most 30 days. If you ask for an earlier start, the Alchemy CLI clamps it to your plan's limit and shows the range it returned.

Can AI coding agents check usage with the Alchemy CLI?

Yes. Any agent that runs terminal commands, including Claude Code, Cursor, and Codex, can call alchemy --json --no-interactive usage summary and read the JSON result. Running alchemy --json --no-interactive agent-prompt gives the agent the full command tree, auth methods, and error codes, so it calls the Alchemy CLI correctly without trial and error. Agents connected to the Alchemy MCP Server can also use its get_usage_summary and get_usage_time_series tools.

Start checking usage from your terminal

Install the CLI and run alchemy usage summary to see where this month stands. It signs in with the account you already use for the Dashboard, the usage commands need no API key, and they work on every plan. New to Alchemy? Create a free account in the Alchemy Dashboard, then point the CLI at it with alchemy auth.