Connect Scrape.ly with your workflow

When a run finishes, start another Actor or one of your saved tasks. Chain search, detail and clean-up steps without a server.

POST to any URL when a run or build succeeds, fails, aborts or times out. Custom payload and header templates, retries and a delivery log.

Post a message to your Slack channel when a run succeeds, fails or times out, with result count, cost, links and a preview of the first results.

Send run alerts to a Discord channel through a channel webhook. Pick the events, add your own message, preview the first results.

Email a run summary to your account address, or to any address you confirm, when a run finishes. One-click unsubscribe.

Write the results of every successful run into a Google Sheet you share with Scrape.ly. Add rows or replace the tab, with nested fields as columns.

Start Actors and saved tasks on a cron timetable, several at once, with an email alert when a scheduled run fails to start.

Save an input and run options under a name, then run it from the dashboard, the API, a schedule or a webhook.

Create Actors from your repositories. Every push to the linked branch rebuilds the Actor automatically.

Get told about failed schedules, low balance, expiring storage and your Actors, by email or in the dashboard. Low-balance alerts can also call a webhook.

One URL for every MCP client. Agents search the Store, run Actors and read results. Sign in with OAuth or an API key.

Add Scrape.ly as a custom connector in Claude, or with one command in Claude Code. OAuth sign-in, no key to copy.

Connect the hosted server as a custom MCP server in ChatGPT, or add it to Codex from the terminal.

Add the server to mcp.json and let Cursor run scrapers while you code.

Use Scrape.ly tools from Copilot agent mode with a single mcp.json entry.

Point Windsurf at the hosted endpoint and call Actors from Cascade.

Add the remote server in Cline and give your agent live web data.

Register Scrape.ly as a context server in Zed’s settings.

A stdio server on PyPI for clients that can’t reach a remote server. Run it with uvx, choose which tool groups to load.

Actors call any OpenRouter model with their run token, no API key of their own. Usage is billed to the run.

Run Actors and tasks, read datasets, manage schedules and webhooks over HTTP. Existing Actor API clients work by changing the base URL.

Typed API client for Node.js, browsers and edge runtimes, a drop-in replacement for existing Actor API clients.

Sync and async API client for Python. Run Actors and stream results in a few lines.

Build Actors in Node.js on Crawlee: storages, proxies, the hosted browser and pay-per-event charging.

Build Actors in Python with async storages, proxy configuration and events.

Create, run and push Actors from your terminal, manage storages, tasks, schedules, webhooks and secrets, and script it all in CI.

No-code tools work today through their HTTP steps: start runs and fetch results from n8n, Make, Zapier or Pipedream with a Bearer token, and trigger workflows from webhooks.

Trigger a build from any workflow with the Actor’s build URL, or run the CLI in your pipeline.

Download results as JSON, JSONL, CSV, Excel, XML, HTML, RSS or PDF, with field selection, flattening and unwinding.

Share a dataset without your token. Expiring links work in any tool that can fetch a URL, in any export format.

Post run results to a channel today: point an HTTP webhook at a Slack incoming-webhook URL with a payload template.