MrScraper
MrScraper provides web scraping APIs for structured extraction, search results, marketplace data, travel data, and web unblocking.
Verdict
Common use cases
- Pull competitor pricing from e-commerce sites
- Build lead lists from directory pages
- Monitor job postings across career portals
- Extract event details from conference schedules
- Aggregate news headlines for daily briefs
Integration
- Vendor
- MrScraper
- Category
- other
- Auth
- API_KEY
- Tools
- 11
- Composio slug
mrscraper
Tools
- Bulk Rerun AI Scraper
Run an existing General AI scraper configuration against multiple URLs in one request. Returns the accepted bulk result ID for Get Scrape Result.
- Bulk Rerun Manual Scraper
Run an existing saved manual scraper workflow against multiple URLs in one request. Returns the accepted bulk result ID for Get Scrape Result.
- Create AI Scraper
Create and run a general, listing, or map AI scraper against a website.
- Get Account Usage
Return the connected MrScraper account's plan, token usage, concurrency allowance, and rate limits without exposing user identity or token details.
- Get Scrape Analytics Timeline
Return time-series scrape activity and success metrics for a domain and action over a time range.
- Get Scrape Result
Retrieve the current status and output for one accessible scrape result by ID.
- Get Scrape Status Analytics
Return aggregate scrape counts, success rate, request rate, latency, and status totals for a domain and action over a time range.
- Heal Manual Scraper
Ask MrScraper to repair an existing manual scraper workflow that no longer matches its target site.
- List Scrape Results
List scrape results for the connected account with filtering, sorting, and agent-controlled pagination.
- Rerun AI Scraper
Run an existing AI scraper configuration against a new URL, with optional map crawl controls.
- Rerun Manual Scraper
Run an existing saved manual scraper workflow against a new URL.
Setup
Setup guide
- 11. In Switchy, open Settings → Integrations → MCP Servers and click Add Server. 2. Select MrScraper from the catalog. 3. You'll be prompted for an API key — log into your MrScraper dashboard, navigate to API Settings, and generate a new key with read and write permissions. 4. Paste the key into Switchy and click Connect. 5. Switchy will verify the connection by calling Get Account Usage to confirm your plan and token balance. 6. Open any Project, type '@MrScraper' in the chat, and ask it to scrape a URL — for example, '@MrScraper extract the product name, price, and availability from https://example.com/shop'. 7. The MCP will create an AI scraper, run it, and return the structured JSON result in the thread. 8. To reuse a scraper, note the result ID and invoke Bulk Rerun AI Scraper with a list of new URLs.
What teammates see: by default, memories from MrScraper are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Scrape Product Catalog
@MrScraper scrape https://example.com/products and return the product name, price, currency, and in-stock flag for every item on the page.
Build Contact Directory
@MrScraper extract all employee names, job titles, email addresses, and phone numbers from https://example.com/team.
Monitor Competitor Pricing
@MrScraper use bulk rerun on scraper ID 12345 with these URLs: https://competitor-a.com/product, https://competitor-b.com/product, https://competitor-c.com/product.
Aggregate Event Listings
@MrScraper scrape https://example.com/events and return the event title, start date, venue, and signup URL for each listing.
Check Scraper Health
@MrScraper show me scrape status analytics for example.com over the last 7 days, including success rate and average latency.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@MrScraper create an AI scraper to extract product names, prices, and availability from https://example-store.com/catalog
I've created and run an AI scraper for that catalog page. The scraper extracted 24 products with their names, prices, and stock status. For example: "Wireless Headphones" at $89.99 (In Stock), "USB-C Cable" at $12.49 (Low Stock), and "Laptop Stand" at $45.00 (Out of Stock). The full result set is now available with scrape ID scr_abc123. This configuration is saved and can be rerun against other catalog URLs using the bulk rerun tool.
This example shows MrScraper's core AI extraction capability — you describe what data you want, and it infers the structure from the page. The scraper consumes tokens from your MrScraper plan quota. If the site layout changes significantly, you may need to recreate the scraper or use the heal tool for manual workflows.
@MrScraper bulk rerun my saved product scraper against these 50 competitor URLs and return a CSV of all pricing data
I've queued a bulk scrape job across those 50 URLs using your saved scraper configuration. The job ID is bulk_xyz789. MrScraper will process these URLs respecting your account's concurrency limit (typically 5-10 simultaneous requests). Once complete, I'll retrieve the results and compile them into a CSV showing product names, prices, and URLs. Estimated completion time is 3-5 minutes depending on page load times and your plan's rate limits.
Bulk rerun is ideal for competitive analysis or monitoring workflows. Be aware that 50 URLs will consume 50 scrape credits from your plan. The concurrency limit means large batches take time to complete. If any URLs fail (404, timeout, bot detection), those rows will be marked as errors in the final result set.
@MrScraper show me scrape analytics for example-store.com over the last 7 days, then summarize which pages are failing most often
I pulled the analytics timeline for example-store.com. Over the past week, you ran 342 scrapes with an 87% success rate. The status breakdown shows 298 successful, 31 timeouts, and 13 blocked requests. The highest failure rate occurred on /search pages (22% timeout rate) and /product/detail pages with dynamic lazy-loading (18% block rate). Peak scrape activity was Tuesday at 2 PM UTC. Based on this, the search results page may need a longer wait time in your scraper config, and product detail pages might benefit from a manual workflow with explicit scroll triggers.
This example pairs MrScraper's analytics API with AI reasoning to surface actionable insights. The analytics tools don't expose raw HTML or scrape content — only metadata like status codes, latency, and success rates. Use this to diagnose scraper health before burning through your token budget on broken workflows.
Use-case deep-dives
When bulk scraping beats manual price checks for e-commerce teams
A 6-person e-commerce team tracking 200 competitor SKUs weekly used to split the work across spreadsheets and browser tabs. MrScraper's bulk rerun tools let them configure one AI scraper per competitor site, then fire all 200 URLs in a single request each Monday morning. The Get Scrape Result tool polls for completion, and the team has a clean dataset in under 10 minutes. This works when competitor sites share similar structure (product grids, price tags in predictable spots). If you're scraping sites that redesign monthly or use heavy JavaScript rendering, the Heal Manual Scraper tool becomes essential—but that adds a maintenance tax. For stable competitor catalogs under 500 URLs per run, this MCP turns a half-day chore into a 10-minute standup agenda item.
Why this MCP struggles with real-time prospecting workflows
A 3-person sales team wanted to scrape LinkedIn profiles and company pages to enrich leads during discovery calls. MrScraper's Create AI Scraper can extract bio text and job titles, but the async result model (submit job, poll Get Scrape Result, parse JSON) adds 15-30 seconds of latency per lookup. That's too slow when a prospect is on the phone and the rep needs enrichment data now. The Get Account Usage tool shows token burn and rate limits, which matter when you're scraping hundreds of profiles daily, but the MCP doesn't expose webhooks or streaming results. If your workflow is batch-oriented (enrich 50 leads overnight before a campaign launch), this works. For synchronous, in-call enrichment, you need a different tool or a pre-warmed cache.
When AI scrapers save hours on weekly roundup research
A 2-person content team publishes a weekly industry newsletter sourced from 30 blogs and news sites. They used to skim each site manually, copy-paste headlines, and tag themes in Notion. MrScraper's listing AI scraper pulls article titles, publish dates, and summaries from each source in one bulk request. The Get Scrape Analytics Timeline tool tracks which sites break most often (redesigns, paywalls), so the team knows where to invest in manual scraper healing. The workflow runs Sunday night; Monday morning the team reviews 200 pre-extracted items instead of visiting 30 sites. This breaks down if sources require login (auth isn't exposed through the MCP) or if you need sub-article analysis like sentiment scoring. For public content aggregation at 20-50 sources, this MCP cuts research time by 70 percent.
Frequently asked
What does the MrScraper MCP do in Switchy?
It lets your team extract structured data from websites without writing scraper code. You can create AI-powered scrapers that adapt to page layouts, rerun saved workflows across multiple URLs, and retrieve results directly in your Switchy workspace. The MCP handles both one-off scrapes and bulk operations, returning JSON data you can feed into other tools or prompts.
Do I need a paid MrScraper account to use this MCP?
Yes. You connect via API key, which means you need an active MrScraper subscription. The Get Account Usage tool shows your plan limits and token balance, so Switchy can warn you before you hit rate caps. Free-tier keys work, but bulk scraping and high concurrency require a paid plan on MrScraper's side.
Can the MCP scrape sites that require login or JavaScript rendering?
MrScraper handles JavaScript-heavy sites, but authentication depends on what the vendor supports in their API. The MCP itself doesn't manage cookies or session state—you'd configure auth inside MrScraper's dashboard first, then reference that scraper config by ID. For sites with CAPTCHAs or aggressive bot detection, results vary and you may need MrScraper's premium proxy features.
Why use this instead of calling MrScraper's API directly?
The MCP wraps the API in natural language. Instead of constructing JSON payloads and polling endpoints, your team asks Switchy to scrape a URL or check scrape status, and the assistant handles the request lifecycle. You also get inline usage warnings and error explanations, which raw API calls don't provide. If you're already scripting against the API, the MCP adds little value.
Who on the team should connect the MrScraper account?
Whoever owns your MrScraper subscription and can generate an API key. That person's token usage and rate limits apply to every scrape Switchy runs, so pick someone with a plan tier that matches your team's volume. Other workspace members don't need their own MrScraper accounts—they share the connected integration and its quota.