Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

We built a time machine for the web. Introducing Exa Snapshot: an index of 400 billion historical snapshots of webpages that lets you search as if it's the past. Snapshot is already being used for backtesting prediction models, RL at labs, exploring the pre-AI web, and more.

Exa

62,548 subscribers

647,035 views • 1 day ago •via X (Twitter)

News & Politics Education Science & Technology

Anya Rossi• Live Now

Private livecam show

32 Comments

Exa1 day ago

Snapshots are available to try in Exa today. Learn more:

Ishan Goswami1 day ago

the internal codename for this project was...

Mridul Singhai1 day ago

How does this differ from the venerable Wayback Machine from Internet Archive?

jacky1 day ago

out of curiosity, how is this different from the internet archive's version?

Max Rovensky1 day ago

You built

iFred1 day ago

How much did you folks donate to the Internet Archive?

Jakub Hojsan1 day ago

wait the audio kinda cooked

Jessica1 day ago

so cool seeing landing page taglines over the years

Steven Liss1 day ago

feature request: let me search the web 1 day/month/year in the future

&.1 day ago

this is very cool

Bobir M1 day ago

What is difference from waybackmachine ?

Jed White 💥♻️1 day ago

Congrats and cool effort! Storage must be wild 🫡

khang1 day ago

wow

Aapakari1 day ago

How often are pages recaptured?

r/691 day ago

so you reinvented archive[.]org

Mustafa1 day ago

Only 2 eras Chat, pre-chat

Aniket1 day ago

the underrated use case: checking what a competitor's pricing page said last quarter

Mudit Arora1 day ago

glad we’re using you as our service

Phil Gara1 day ago

Wow really cool!

Akul Gupta1 day ago

Super cool, gonna be super useful for research

Seth BRKV1 day ago

No way! I thought about such system yesterday morning. How you can test AI inside of "real" internet, but months or years behind. And there it is (if I understood correctly)

.CitiZen.1 day ago

Woah

Rimas1 day ago

how are labs using the snapshots for rl?

echo4eva1 day ago

so cool!

Harjass Gambhir1 day ago

is there a filter for pre gpt pages for searching?

Krish Goel1 day ago

This was a major bottleneck when my friends and I (naively) tried to build a market prediction engine that needed to understand the market sentiment “then”. Love it.

AI News Daily1 day ago

A historical web index is a surprisingly useful primitive for AI research. Backtesting agents and models against old pages could make evaluations much closer to the messy conditions they face in the wild.

Dozer🚜1 day ago

Wow dope idea

RISHI1 day ago

👀👀👀

Callum1 day ago

super nice for evals!

Ryan1 day ago

400b snapshots turns search from "what does this page say?" into "when did this claim first appear?" way more useful for due diligence than another current-web wrapper.

Jason Garoutte1 day ago

AI is fundamentally about predictions. Having date-specific data is key.

Related Videos

🔍 Introducing BackSearch. LLMs are increasingly asked to predict the future, but a good backtest requires a snapshot of the internet at a point in time. BackSearch allows LLMs to search the web as it was on a particular date. It’s great for: 🔮 Forecasting and prediction markets. 📈 Quantitative finance. 🌍 RL environments that simulate the world. 🥶 Freezing websearch for benchmark reproducibility We’re releasing a narrow slice of our index to begin with, focused on the news domain for 2026. Based on feedback we’ll open up more of our index in subsequent releases. 👇 Try BackSearch out with the link below.

🔍 Introducing BackSearch. LLMs are increasingly asked to predict the future, but a good backtest requires a snapshot of the internet at a point in time. BackSearch allows LLMs to search the web as it was on a particular date. It’s great for: 🔮 Forecasting and prediction markets. 📈 Quantitative finance. 🌍 RL environments that simulate the world. 🥶 Freezing websearch for benchmark reproducibility We’re releasing a narrow slice of our index to begin with, focused on the news domain for 2026. Based on feedback we’ll open up more of our index in subsequent releases. 👇 Try BackSearch out with the link below.

General Reasoning

419,098 views • 1 month ago

I think we've been approaching AI web search the wrong way. Instead of searching for information... I tried assigning an entire research task. I built a Web Search Agent to track the biggest AI launches from the past week. Here's what happened.

I think we've been approaching AI web search the wrong way. Instead of searching for information... I tried assigning an entire research task. I built a Web Search Agent to track the biggest AI launches from the past week. Here's what happened.

Kaitee

58,149 views • 1 month ago

Fable 5 is so good at web design that I genuinely don't see how web design careers make it past the end of this year. An AI that can one shot designs this good means that the web design industry as a whole is in serious trouble. If you're in web design, it's time to pivot.

Fable 5 is so good at web design that I genuinely don't see how web design careers make it past the end of this year. An AI that can one shot designs this good means that the web design industry as a whole is in serious trouble. If you're in web design, it's time to pivot.

Mark Gadala-Maria

543,650 views • 2 months ago

Today we are announcing Keenable AI, an AI-native index of the best human knowledge we have, starting with the open web. AI seems to know just about everything until you ask it about something you know deeply. The answer isn’t wrong, but it's just very average. We started Keenable to fix this problem: every model and every agent should be able to query, reason over, and continuously learn from the living web. Backed by a $26M Seed from Accel and Conviction. Built by the team that took on Google at Yandex Search, with researchers and engineers from Amazon AGI, X, and Perplexity. The Web Search API and Web Query Language for AI are live now. Together they make the web queryable at AI scale. Both are free until end of September. Seven days of benchmarks, stories, and launches ahead. Let your agents search.

Today we are announcing Keenable AI, an AI-native index of the best human knowledge we have, starting with the open web. AI seems to know just about everything until you ask it about something you know deeply. The answer isn’t wrong, but it's just very average. We started Keenable to fix this problem: every model and every agent should be able to query, reason over, and continuously learn from the living web. Backed by a $26M Seed from Accel and Conviction. Built by the team that took on Google at Yandex Search, with researchers and engineers from Amazon AGI, X, and Perplexity. The Web Search API and Web Query Language for AI are live now. Together they make the web queryable at AI scale. Both are free until end of September. Seven days of benchmarks, stories, and launches ahead. Let your agents search.

Andrey Styskin

549,804 views • 24 days ago

🚨 THIS IS ACTUALLY INSANE Your AI agent can have access to the web. But if it can't reliably read what’s actually on the page, that access is almost useless. We looked at Firecrawl as the web layer for AI agents and the numbers are hard to ignore. The setup is simple: Give it a URL, search query, or website. Firecrawl handles the ugly part — crawling, scraping, rendering, extracting, and turning web content into something an AI model can actually use. The headline numbers: → 173,000+ GitHub stars → Search, scrape and interact with the web at scale → Supports web pages, PDFs, DOCX and other content → Structured data extraction for AI workflows → MCP support for connecting it directly to AI agents The workflow looks like this: Search → Scrape → Crawl → Extract → Feed the agent Three things stand out: 1. Scraping becomes an infrastructure layer Instead of maintaining your own pile of HTTP clients, parsers, browser automation and retry logic, you can treat web access as an API. 2. Agents get more than raw HTML The goal isn't just downloading a webpage. It's turning messy web content into clean context that an LLM can reason over. 3. The same layer works across different agent workflows Research agents. RAG pipelines. AI search. Competitive intelligence. Web-data extraction. The interesting shift: AI agents don't just need better models. They need better access to the information those models are supposed to reason about. Firecrawl is building that layer. Save this repo.

🚨 THIS IS ACTUALLY INSANE Your AI agent can have access to the web. But if it can't reliably read what’s actually on the page, that access is almost useless. We looked at Firecrawl as the web layer for AI agents and the numbers are hard to ignore. The setup is simple: Give it a URL, search query, or website. Firecrawl handles the ugly part — crawling, scraping, rendering, extracting, and turning web content into something an AI model can actually use. The headline numbers: → 173,000+ GitHub stars → Search, scrape and interact with the web at scale → Supports web pages, PDFs, DOCX and other content → Structured data extraction for AI workflows → MCP support for connecting it directly to AI agents The workflow looks like this: Search → Scrape → Crawl → Extract → Feed the agent Three things stand out: 1. Scraping becomes an infrastructure layer Instead of maintaining your own pile of HTTP clients, parsers, browser automation and retry logic, you can treat web access as an API. 2. Agents get more than raw HTML The goal isn't just downloading a webpage. It's turning messy web content into clean context that an LLM can reason over. 3. The same layer works across different agent workflows Research agents. RAG pipelines. AI search. Competitive intelligence. Web-data extraction. The interesting shift: AI agents don't just need better models. They need better access to the information those models are supposed to reason about. Firecrawl is building that layer. Save this repo.

Vikas gupta

18,053 views • 13 days ago

The internet wasn't built for AI. We're fixing that. Introducing Scout by Chakra - map the web as you browse. Climb the leaderboard and help turn the internet into an open, permissionless database.

The internet wasn't built for AI. We're fixing that. Introducing Scout by Chakra - map the web as you browse. Climb the leaderboard and help turn the internet into an open, permissionless database.

Chakra Labs

109,760 views • 1 year ago

No Three.js. No fake web-based depth-of-field or focal-length emulation. Customize in web, render in Blender. While everyone is vibe-coding apps for flat pattern generators, or at most Three.js-based 3D scenes, we’re already one step ahead. We built an app that lets you use a web UI to control an actual Blender 3D render. Eevee engine is used for streaming quick previews back into the app, while Cycles renders the final image for better material quality and lighting. So simple to use that you can even give it to your client. Of course, built with Toolcraft.

No Three.js. No fake web-based depth-of-field or focal-length emulation. Customize in web, render in Blender. While everyone is vibe-coding apps for flat pattern generators, or at most Three.js-based 3D scenes, we’re already one step ahead. We built an app that lets you use a web UI to control an actual Blender 3D render. Eevee engine is used for streaming quick previews back into the app, while Cycles renders the final image for better material quality and lighting. So simple to use that you can even give it to your client. Of course, built with Toolcraft.

Alex Barashkov

76,542 views • 2 months ago

What if AI chat is our new front door to the web? Meet "Inline Browsing" in Dia (available today) Now Dia opens webpages *within* your AI chat threads – blurring the lines between 3 categories of software: a web browser, search engine, and AI chat. Let me explain...

What if AI chat is our new front door to the web? Meet "Inline Browsing" in Dia (available today) Now Dia opens webpages within your AI chat threads – blurring the lines between 3 categories of software: a web browser, search engine, and AI chat. Let me explain...

Josh Miller

74,888 views • 1 year ago

The paradigm shift is here: We are moving from a "Human-First" internet to an "Agent-First" internet. 🤖 In the latest episode of the AI on Air podcast, our CTO Scott Shi - e/acc dives deep into the future of the Agent-to-Agent (A2A) economy. Imagine a world where your AI assistant doesn't just search the web, but autonomously hires and pays other specialized agents to complete complex tasks for you. The infrastructure for this autonomous machine economy is being built right now. Are you ready for the future of AI payments? Watch the full episode to learn more⬇️

The paradigm shift is here: We are moving from a "Human-First" internet to an "Agent-First" internet. 🤖 In the latest episode of the AI on Air podcast, our CTO Scott Shi - e/acc dives deep into the future of the Agent-to-Agent (A2A) economy. Imagine a world where your AI assistant doesn't just search the web, but autonomously hires and pays other specialized agents to complete complex tasks for you. The infrastructure for this autonomous machine economy is being built right now. Are you ready for the future of AI payments? Watch the full episode to learn more⬇️

KITE AI

25,038 views • 6 months ago

Today I'm launching Muse It's an AI tool for music producers that lets you generate MIDI and synth presets, available for Ableton Live and on the web.

Today I'm launching Muse It's an AI tool for music producers that lets you generate MIDI and synth presets, available for Ableton Live and on the web.

Jake

68,043 views • 1 year ago

We explored the use of 3D GS for museums, using an MBAM gallery as our testbed. A single on-site capture rendered across web, VR, and installation. Thank you to the MBAM for the support. See & try: Here is the web version built using PlayCanvas

We explored the use of 3D GS for museums, using an MBAM gallery as our testbed. A single on-site capture rendered across web, VR, and installation. Thank you to the MBAM for the support. See & try: Here is the web version built using PlayCanvas

Hugues Bruyère

16,557 views • 1 year ago

Introducing Onboard Business 💼 We started Onboard for professionals in Africa and the Global South who needed direct access to the global economy The more we built for consumers, the more we saw businesses dealing with the exact same problems on a bigger scale So we built Onboard Business. Same infrastructure our consumers already trust, now as a web app for businesses

Introducing Onboard Business 💼 We started Onboard for professionals in Africa and the Global South who needed direct access to the global economy The more we built for consumers, the more we saw businesses dealing with the exact same problems on a bigger scale So we built Onboard Business. Same infrastructure our consumers already trust, now as a web app for businesses

Yele | Onboard 🌐

11,732 views • 2 months ago

I've been working on web projects with Three.js more and more lately — 2D, 3D, XR — and the one thing I've noticed is the lack of tools compared to game engines. That's why I decided to start building on the side a library of 3D assets that are actually little programs, AI-native, and fully customizable. I called this @polyfork_ , and it's already being used by dozens of people that have an AI centric workflow, since the website can be fully understood and used by agents. Assets are being built each day, at a rate of 100+ a week, and I'm planning to make it the biggest programmatic-3D library on the internet. Almost half of the models are free, and I opened slots for "founding" members, who can make a one-time payment for lifetime access, with access to all current models and kits, plus everything built from now on. Get one before they run out.

I've been working on web projects with Three.js more and more lately — 2D, 3D, XR — and the one thing I've noticed is the lack of tools compared to game engines. That's why I decided to start building on the side a library of 3D assets that are actually little programs, AI-native, and fully customizable. I called this @polyfork_ , and it's already being used by dozens of people that have an AI centric workflow, since the website can be fully understood and used by agents. Assets are being built each day, at a rate of 100+ a week, and I'm planning to make it the biggest programmatic-3D library on the internet. Almost half of the models are free, and I opened slots for "founding" members, who can make a one-time payment for lifetime access, with access to all current models and kits, plus everything built from now on. Get one before they run out.

Lucas Martinic

35,302 views • 1 month ago

Introducing Stora, AI agents for the app store The app store is a full-time job nobody signed up for Screenshots. Compliance. ASO. Publishing We built agents to take it all on Shipping on mobile is now as easy as shipping to web Stora |

Introducing Stora, AI agents for the app store The app store is a full-time job nobody signed up for Screenshots. Compliance. ASO. Publishing We built agents to take it all on Shipping on mobile is now as easy as shipping to web Stora |

Carlton Aikins

13,883 views • 5 months ago

This is the DeepSeek moment of Search? A small team of cracked engineers built a search API for AI agents that outperforms everything built with 20x the funding. AI systems are about to become the biggest users of information on Earth. They don’t need a better Google. They need a retrieval layer built for reasoning, scale, and access beyond the web. That’s what we’re building. 1/3

This is the DeepSeek moment of Search? A small team of cracked engineers built a search API for AI agents that outperforms everything built with 20x the funding. AI systems are about to become the biggest users of information on Earth. They don’t need a better Google. They need a retrieval layer built for reasoning, scale, and access beyond the web. That’s what we’re building. 1/3

Valyu

1,135,072 views • 11 months ago

it's awake. the way you interact with the web, information, and services is being rewritten. introducing FlowithOS — the world's first operating system natively built for ai agents. self-evolving. memory-powered. lightning-fast. beyond any ai browser, it's the SMARTEST agentic os that turns your browser into real-world value, from assisting you to acting for you. let's witness together ⬇️

it's awake. the way you interact with the web, information, and services is being rewritten. introducing FlowithOS — the world's first operating system natively built for ai agents. self-evolving. memory-powered. lightning-fast. beyond any ai browser, it's the SMARTEST agentic os that turns your browser into real-world value, from assisting you to acting for you. let's witness together ⬇️

Flowith

1,285,219 views • 10 months ago

🚨NEW LABS EXPERIMENT 🚨 Introducing Disco, an experimental way to discover new generative AI features on the web. Our first feature, GenTabs, uses Gemini 3 to remix your open tabs into totally custom apps to help you get more out of the web. Learn more and join the waitlist →

🚨NEW LABS EXPERIMENT 🚨 Introducing Disco, an experimental way to discover new generative AI features on the web. Our first feature, GenTabs, uses Gemini 3 to remix your open tabs into totally custom apps to help you get more out of the web. Learn more and join the waitlist →

Google Labs

1,761,288 views • 9 months ago

Elon Musk: You might be able to download a snapshot of your mind to an Optimus robot within 20 years. Question from the audience: “Do you see a path for Optimus to have human consciousness downloaded to it?” Elon: “It's not immediate, but if you say, down the road, would you be able to, say, with a Neuralink, have a snapshot of what is an approximate snapshot of somebody's mind and then upload that approximate snapshot to an Optimus body? I think that at some point that technology becomes possible, and it's probably less than 20 years. Of course you won't quite be the same; you'll be a little different because you'll be in a robot body, and the mental snapshot will not be precise. It'll probably be pretty close, but not exactly the same. On the other hand, are you the same person that you were five years ago? Nope. I mean a lot of things have changed, so. Yeah. But I guess at some point if you want to be uploaded to a robot body, my guess is that becomes possible.” Tesla Shareholder Meeting, November 6, 2025

Elon Musk: You might be able to download a snapshot of your mind to an Optimus robot within 20 years. Question from the audience: “Do you see a path for Optimus to have human consciousness downloaded to it?” Elon: “It's not immediate, but if you say, down the road, would you be able to, say, with a Neuralink, have a snapshot of what is an approximate snapshot of somebody's mind and then upload that approximate snapshot to an Optimus body? I think that at some point that technology becomes possible, and it's probably less than 20 years. Of course you won't quite be the same; you'll be a little different because you'll be in a robot body, and the mental snapshot will not be precise. It'll probably be pretty close, but not exactly the same. On the other hand, are you the same person that you were five years ago? Nope. I mean a lot of things have changed, so. Yeah. But I guess at some point if you want to be uploaded to a robot body, my guess is that becomes possible.” Tesla Shareholder Meeting, November 6, 2025

ELON CLIPS

49,486 views • 10 months ago