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

# Daemon Mode

> Run Reader as a long-lived background process with a shared browser pool.

Every time you invoke `reader scrape` or `reader crawl`, Reader normally spins up a fresh browser pool, runs your command, and tears it down. That's fine for one-off scripts but wasteful when you're making many calls - you pay the 1-2 second pool startup cost every time.

Daemon mode fixes this. Start a background daemon once, and all subsequent CLI commands auto-attach to it, reusing the warm pool.

## Start the daemon

```bash theme={null}
reader start --pool-size 5
```

This starts a background process listening on port 4000 (default) with a 5-instance browser pool. The first CLI command will feel instant - no startup cost.

Optional flags:

| Flag              | Purpose                          |
| ----------------- | -------------------------------- |
| `--pool-size <n>` | Browser pool size (default: 5)   |
| `--port <n>`      | Listen port (default: 4000)      |
| `--show-chrome`   | Show browser windows (debugging) |
| `-v, --verbose`   | Enable logging                   |

## Check status

```bash theme={null}
reader status
```

Reports whether the daemon is running, its pool size, and the number of in-flight requests.

## Use the daemon transparently

Once the daemon is running, just use `reader scrape` and `reader crawl` as normal:

```bash theme={null}
# First call - attaches to daemon automatically
reader scrape https://example.com

# Second call - also uses the daemon, instant startup
reader scrape https://another.com
```

The CLI auto-detects a running daemon on the default port. To use a custom port:

```bash theme={null}
reader start --port 5000
reader scrape -p 5000 https://example.com
```

## Bypass the daemon

For one-off runs where you don't want to touch the daemon (e.g., you want to test without affecting the shared pool):

```bash theme={null}
reader scrape --standalone https://example.com
```

This runs in standalone mode regardless of whether a daemon is running.

## Stop the daemon

```bash theme={null}
reader stop
```

## When to use daemon mode

* **Shell scripts** that run Reader many times in sequence
* **Development** where you're iterating on CLI commands
* **Cron jobs** that fire frequently (keep the daemon warm rather than cold-start on every run)

## When not to use it

* **Single one-off commands** - the daemon startup cost exceeds the savings
* **CI/CD** - each build should be stateless; just use `--standalone` or skip daemon entirely
* **Server applications** - embed `ReaderClient` directly in your Node app instead; the CLI daemon is for terminal workflows

## Where to go next

<CardGroup cols={2}>
  <Card title="CLI guide" icon="terminal" href="/self-hosted/guides/cli">
    Commands and flags for scrape and crawl.
  </Card>

  <Card title="Deployment" icon="server" href="/self-hosted/guides/deployment">
    Running Reader as a production service.
  </Card>
</CardGroup>