Misha Zanka (@Raznem) profile picture

Building scraping agents at https://teapawt.net/t.co/kZqUDLGlhj Co-Founder of https://teapawt.net/t.co/Uscbiyw7N3

Joined February 2012
Photo by Misha Zanka (@Raznem)Photo by Misha Zanka (@Raznem)Photo by Misha Zanka (@Raznem)Photo by Misha Zanka (@Raznem)
Thanks, Fable. Now I finally touched Claude's sub limit
32
If you’re building AI agents but are not doing automation, you’re completely missing out. As a developer of a viral AI investor agent (@nocap_so), I’m convinced that once it works, it would be the long-awaited 100x unlock of the whole ecosystem. And we have just released the perfect resource to get started - “State of Browser/Computer Use 2025” report Inside: - a comprehensive market map of 60+ companies across developer (@browserbase, @browser_use, @skyvernai, @onkernel, @hyperbrowser, @joinmassive, @steeldotdev, @HalluminateAI, @apify, @ExaAILabs, @firecrawl, @IntunedHQ); research (@AnthropicAI, @induction_labs, @OpenAI) & consumer (@GeneralAgentsCo, @diabrowser, @strawberry, @comet, @yutori_ai) toolstack - a practical playbook for how to navigate it - a list of future benchmarks/leaderboards in the making you can contribute to! The best part - it's 100% free with no gate-keeping or begging for "comments" - link is the the thread! As far as we're concerned, this is the most up-to-date map in the space. This is just a start - expect more analysis & benchmarks to come in the coming months. 📍 Bookmark this post for later so you don't lose it.
9
7
5
39
6,013
5% battery. One live demo with @AemalSayer. Parsera’s n8n node. A wild bug appears — then vanishes for hours. Things went wrong at "the perfect timing"
1
48
For the full version of the pod, look here: youtube.com/watch?v=aXMLuPtA…
34
After 40 minutes of being down, finally @Google admitted (though partially) that it's down as half of the internet (which probably uses google cloud as well)
1
8
45
12,413
The LLM capability mess is over thanks to Parsera's @raznem! 🎉 The API I announced last month is now LIVE. Get real-time model info (pricing, context windows, capabilities) for all major LLMs in one place. Blog: paolino.me/standard-api-llm-… Model browser: llmspecs.parsera.org/
1
3
1
17
962
Misha Zanka retweeted
Parsera is now Actorized 📦 One of your favorite scraping libraries on GitHub is now easier to run - directly in the cloud with the Parsera Actor on Apify. Built by the community, integrated with ours. Just input your target URL and get clean, structured data back. Link in 🧵
2
2
8
368
What's wrong with X? Constant clickbait, and now it's progressing to copy-paste, what
67
LLM providers don't offer pricing or capabilities in their list models endpoints. This ends NOW. I'm partnering with parsera.org to provide a standard pricing and capabilities API for LLM devs. Available in RubyLLM soon. Have a look: gist.github.com/crmne/301be1…
3
2
1
29
4,223
Played around with Deepseek-R1 and the fact that a laptop can run a model with better performance than GPT-4o is mindblowing.
133
Cool list with examples of how you are fingerprinted on the web: browserleaks.com/
101
Writing and maintaining web scraping code requires constant updates due to changing HTML structures and complex selectors, which results in brittle code and frequent breakages. Parsera uses LLMs to extract data from websites based on simple language descriptions.
2
8
46
1,735
2 new features of Parsera API arrived: Cookies support - you can send cookies in request and access login-protected pages: docs.parsera.org/api/cookies… Precision mode - improved accuracy by diving deeper into the page components: docs.parsera.org/api/precisi…
47
We are launching our API, providing you with the infrastructure to extract data from any website at scale. You can explore the full API documentation for more details: docs.parsera.org/api/getting…
53
Documentation of Parsera is finally live: docs.parsera.org/, now you can find how can you set up proxy, change model and execute custom playwright during web scraping with Parsera.
98
Last Friday Parsera 0.1.4 was released. Added new stuff from contributors: - Support of @huggingface transformers - New extraction modes: Item for single item extraction and List for extraction labeled lists from the website. Check new features on Github: github.com/raznem/parsera
147
After a few months of experiments with LLMs to quickly extract structured data from websites, I am sharing my code as an open-source Python library: github.com/raznem/parsera Happy to hear your feedback!
2
114