April Product Update 🎉 - Browser Traces – One viewer with Reworkd/Playwright events, console logs, network activity, and CDP events for easier debugging. - Inline Code Diffs – View side-by-side additions and deletions between the new code and the previous version. - Weekly Summary Email – A report with key metrics: total output counts and the sources with the largest data swings. To learn more about the product update, visit our blog (link in the 🧵).

March Product Update 🚀 We’ve officially launched our self-serve scraping platform — now anyone can try it out. 👉 Try the interactive demo (in the 🧵) What’s new: - Chat-based agent for easier debugging - Real-time streaming – no more page refreshes To learn more about the product update, visit our blog.

How Lookbk Extracted 350,000+ E-Commerce Products with Reworkd Lookbk came to us because they were spending 40+ hours every month fixing their existing web scrapers when websites changed. With plans to scale their data pipeline 10×, they needed a solution that could keep up. "Before Reworkd, we spent countless frustrating hours fixing scrapers every time a site changed—now it’s automated. Their advanced captcha solving also unlocks data from sites we couldn’t access before. Scaling our data pipeline is suddenly no longer an issue." - @caelin_sutch, Co-founder of Lookbk 👇 Check the link in the 🧵 to read the full case study.

January Product Update 🚀 We've rolled out several major updates over the past month: 📋 Review Flow: A built-in review system for the QA team—verify scraped data directly on the platform with full history. 🛡️ Anti-Bot Solution: Improved browser stealth capabilities to bypass even the most heavily protected sites. ⚡ Performance Boost: Increased the platform’s ability to handle an order of magnitude more load per day. We are launching our self‑serve tool on March 11th—DM us or email at srijan@reworkd.ai for early access. To learn more about the product update, visit our blog (link in the 🧵).

How @Axis_HQ Automated Regulatory Data Scraping from 2,500+ Sites with Reworkd Axis came to us after their previous vendor couldn't keep up with their data extraction needs. Scaling to 2,500+ sites seemed daunting - but in just a few weeks, we automated the entire process, helping them extract over 5M+ data points. "The Reworkd team is extremely responsive, and their fully managed solution means we don’t have to worry about the quality assurance of our data. Combine that with their competitive pricing, and it was a no-brainer for us." - Mishaal Al Gergawi, CEO of Axis 👇 Check out the full case study here: reworkd.ai/blog/how-axis-aut…

3. Exports Easily track and manage all of your exported data from our new exports page. - Select the exact site(s) you want data from. - Export the data in your preferred format - JSON or CSV. - Pick a date to export data scraped after that day. - Access export history and more

2. Templates Create reusable code templates for websites that share the same hosting platform, rather than writing separate extraction scripts for each site.

1. Dashboarding A critical part of managing web data workloads is having a holistic understanding of the pipeline: how much data is coming in and where failures are arising. With our dashboard, you can get a comprehensive view of your web data pipeline in one place.

We've rolled out many major updates the past few months including: - a brand-new dashboard page - code templating functionality - and much-improved exporting options. Check out the thread 🧵 below for more info:

📢 Tarsier has grown to >1,200 stars💫—thanks for all the support and contributions! Customers are loving our platform too, with over 100k rows extracted from across the web every day. And this month, our agents 🤖 generated 500+ extractors, i.e. ~15k lines running in prod‼️

It's been a while since our last roundup 🚀 A key metric not listed: +$400 increase in testing costs a day 😅

Excited to start the first issue of the Reworkd roundup! It's been an exciting few weeks for us as we've: - Onboarded customers to production for our platform for web data extraction - Received over 400 stars on GitHub across OSS ⭐ - Created annotations for over 200 evals 🍌

We're a bit late to the party but still super excited to announce that AgentGPT has hit over 25,000 stars on GitHub! ⭐ It feels like it was only yesterday that we pressed publish on our first launch announcement. Thank you everyone for your continued support!

Last month we open sourced two libraries for AI Agents on web tasks. We're excited to announce that these libraries have already amassed a total of 400+ collective stars! Check out github.com/reworkd to find Tarsier and Bananalyzer and try them out today.

😉 agentgpt.reworkd.ai/agent?id…

🙏 agentgpt.reworkd.ai/agent?id…

Not your average blob 🔥 A cool clustering of #AgentGPT's usage through @nomic_ai Atlas

New blob just dropped 🔥

Option 4: Adam's Orb 🫠

Option 3: Evil Orb 😈

Option 2: Lightning Orb ⚡️

Option 1: Disco Orb🪩

Super excited to announce that we are now backed by both @ycombinator and @PanacheVC for our pre-seed round 🎉 If you or anyone you know is keen on leveraging AI agents to automate core business workflows, please sign up for our waitlist (link below)😁 6h6bquxo5g1.typeform.com/to/…

4. Deprecations for existing models 🗑️ If you're hard coding non-stable model names, make sure to update them based on the table below

🚨 Attention devs 🚨 We've updated our setup scripts to be compatible with our new backend setup and also added an interactive CLI ✨ Try it now at github.com/reworkd/AgentGPT!