Pinned Tweet
Introducing Expand Fetch, a new way to scrape and crawl the web.
Your agent has never seen the web. No wonder it gets basic questions wrong.
Traditional Markdown APIs lose critical details. So we trained two System one models to change that:
Iris, a vision model that sees pages like you do, preserving tables, icons, and layouts.
X-Ray, a JSON model that uncovers the hidden data behind every page.
Every fetch also comes with a Wayback Machine. Every line links back to its source, so you can verify exactly what your agent saw.
Three products. One API.
On our benchmark of 1,094 questions across 326 challenging web pages, agents using Expand got 74% right. The next best API got 55%.
Available today at $1 per 1,000 fetches.
Try it at expand.ai
The generated email structure of inbound sales emails is usually this:
[AI summary about Expand]
[Random unrelated pitch]
You can just leave out the AI summary - I already know what we're working on!
Believe it or not, I get automated AI emails of people congratulating on the launch and offering a 90s launch video. Now I'm so curious what I would get with them ...
Thank you everyone for the love and support yesterday. The bottleneck isn’t reasoning but perception.
It was over a year of hard work to make this world wild web work well.
This is the foundation for much more to come.
We‘ll have something to announce soon 🔜
Introducing Expand Fetch, a new way to scrape and crawl the web.
Your agent has never seen the web. No wonder it gets basic questions wrong.
Traditional Markdown APIs lose critical details. So we trained two System one models to change that:
Iris, a vision model that sees pages like you do, preserving tables, icons, and layouts.
X-Ray, a JSON model that uncovers the hidden data behind every page.
Every fetch also comes with a Wayback Machine. Every line links back to its source, so you can verify exactly what your agent saw.
Three products. One API.
On our benchmark of 1,094 questions across 326 challenging web pages, agents using Expand got 74% right. The next best API got 55%.
Available today at $1 per 1,000 fetches.
Try it at expand.ai
2 more things.
1) Read more at expand.ai/blog/announcing-ex…
2) This wouldn't have been possible without the support of some amazing folks, including @ycombinator @rauchg @alanaagoyal @SystemOneVC @swyx
Tim Suchanek retweeted
We've been doing this for centuries in Europe
It's thanks to a technology that the US hasn't yet discovered called "trains"
My goodness, the decaf from frequentcoffee.com/ is the absolute best I ever had.
My favorites so far are frequentcoffee.com/products/…
and frequentcoffee.com/products/…
I don't know how they get so much taste into a decaf. It's incredible.
If you don't do it yet, please start using orchestrator agents. It really changes the game.
You should focus on getting architecture and guardrails right, not scheduling individual chats.
It was a pleasure to experience the release in-person at the Effect meetup in SF! The Effect community is very very talented. So many great projects, like Alchemy.
I'm excited about the improvements. Huuuge step forward!
Effect v4 is here.
One ecosystem. Zero dependencies.
The next chapter of Effect and the foundation for building reliable software and AI agents in TypeScript.
ALT Effect 4.0 stable release announcement on a dark background. Large white “4.0” above “Production-ready. Out now.” A timeline along the bottom progresses from 3.0 Stable to 4.0 RC to 4.0 Stable.
It's crazy to see how much better models just got in the last 4 weeks. I'm coming back to some problems I didn't have time to wrap up like optimizing our K8S cluster, and Opus 5.5 is so incredibly good at visualizing how different placement can look like, it's not even close.
Tim Suchanek retweeted
We're entering a new phase of AI psychosis - factories.
To visualize it, compare these two numbers:
1. Gross developer productivity: how much code you ship
2. Net developer productivity: how much code you ship EXCLUDING WORK ON YOUR FACTORY
They Vary header is very powerful. It's what a big part of Stellate was built on blog.cloudflare.com/vary-sup…
Feedback for @OpenAI : When I start a chat, say with pro reasoning in the iOS app and continue in desktop, it doesn't remember the reasoning level and I have to reset it every time.