Backend & export demos¶

Four tasks for outbound HTTP (no browser) and structured data export.

http-get¶

Fetch a URL via Go's HTTP client — no Chrome window needed.

curl -s -X POST localhost:8765/tasks/backend/http-get -d '{}'
curl -s -X POST localhost:8765/tasks/backend/http-get \
  -d '{"url":"https://api.github.com/repos/chromedp/chromedp"}'

Concepts: http-get, mixing browser and non-browser steps in one bundle.

Use when you've discovered a JSON API via Network → capture.

http-post¶

POST JSON to an endpoint and return the response.

curl -s -X POST localhost:8765/tasks/backend/http-post -d '{}'

Pattern

http-post response url "https://httpbin.org/post" body "{\"key\":\"{{value}}\"}"

Concepts: webhook triggers, API integration without browser overhead.

export-csv¶

Extract data and export as CSV.

curl -s -X POST localhost:8765/tasks/backend/export-csv -d '{}'

Concepts: export csv, server-side file generation.

export-formats¶

Export the same data as CSV, NDJSON, and Markdown in one task.

curl -s -X POST localhost:8765/tasks/backend/export-formats -d '{}'

Export step

export ndjson path "/tmp/out.ndjson" data "{{items}}"   # csv | ndjson | md

Concepts: multiple export formats, piping extract results to files.

Browser vs backend steps¶

flowchart LR
    subgraph browser["Needs Chrome"]
        goto
        extract
        click
    end
    subgraph backend["Go HTTP client"]
        http-get
        http-post
        export
    end

    browser -->|"DOM data"| backend

Step type	Pool lease?	Speed
Browser (`goto`, `extract`, …)	Yes	Slower
Backend (`http-get`, `export`)	No*	Fast

*Backend steps still run inside a task that may hold a window — structure tasks to minimize browser time.

What's next?¶

Crawl — extract data to export
Deployment — ship exports in production