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.

Introducing Composer — the first AI Agent for document processing. Get to production-grade accuracy, autonomously in minutes. In our early beta, some teams hit 99% accuracy on complex document tasks in under 10 minutes. Composer is an agent built to optimize schemas the same way a human would (but... way faster). Instead of tuning prompts by hand, you point Composer at your eval set inside Extend. Composer will: - analyze where your schema falls short - propose targeted improvements - run multiple experiments in parallel - surface diffs, accuracy gains, and traces behind each change With this launch, Extend is the only product on the market that helps you reach production-grade accuracy this fast. Composer is live for all Extend customers today! Try it out at the link in comments below.show more

Kushal Byatnal

7,705 subscribers

75,262 views • 1 year ago •via X (Twitter)

Science & Technology News & Politics Education

Anya Rossi• Live Now

Private livecam show

24 Comments

pivot tables1 year ago

Going to assume that since I have no idea what any of this stuff means that I don’t need it.

Nikhil Mishra1 year ago

Hey @kushalbyatnal will this work for complex documents with contextual information like extract pricing from contracts with different pricing configurations? @tryflexprice We have a base version built in house using getmoni-ai/zerox but would love to improve it massively.

Prashant Bellad1 year ago

How is the PII data handled?

Diet Coke Dev1 year ago

The first (247th) cursor for documents. 99% accuracy (on our narrow test case)

hari raghavan1 year ago

Self-driving document extraction, woo! (I think I might be a paperwork nerd.)

scott belsky1 year ago

awesome

arti1 year ago

Is this using @DSPyOSS ?

Kushal Byatnal1 year ago

@DSPyOSS no we looked into DSPy, but ended up deciding to build it in-house from scratch instead

John Zeratsky1 year ago

Dang, you guys are always a step ahead

Davis Treybig1 year ago

This is so cool!

Appunik Studio1 year ago

This breakdown is super helpful. Thanks for sharing your analysis!

Akash_Agyani1 year ago

Open source or paid?

Potentially viral1 year ago

Super interesting!

Suhrab Khan⚡️1 year ago

Composer could redefine document AI! Near-perfect accuracy in minutes is impressive.

Indra prakash1 year ago

Thanks for sharing. A revolution in document processing is moving from Ctrl+F keyword searches to context-aware searching. It preprocesses documents with layout awareness and hierarchical structure, using LLMs for summarization.Found a similar tool by @AndrewYNg called @LandingAI

Alex Yarosh · AI expert · CEO of AI Studio6 months ago

Genuine question: what happens when document formats change or new edge cases emerge? 99% on beta data is different from 99% on production chaos.

Shahriar Tajbakhsh1 year ago

@grok tell me more.

The Longer Game1 year ago

This is a strong step forward for document automation. Hitting near perfect accuracy that quickly will be a real game changer for teams working with complex data.

Yenkay1 year ago

@grok Provide the differentiators between Composer and Landing AI'S DPT.

Bryan Patricca1 year ago

Include the link, so it's easy to find. Excited to try.

16VC1 year ago

Composer sounds like a total game changer for document processing love how it takes the guesswork out of schema tuning. Curious, what’s been the most surprising improvement teams have seen using Composer so far?

Adam Hardej1 year ago

the crushing continues!!

Zikrillo Akhror-Murod🇺🇿🇵🇸1 year ago

Congrats! Could you share the link?

Jeeva1 year ago

How does it work for spatial data in PDF ? LLMs perform poorly though

Related Videos

So the composer for the game leaked some tracks for the Console version on the internet back in 2022. There’s also an interview the composer I’ll link to below.

So the composer for the game leaked some tracks for the Console version on the internet back in 2022. There’s also an interview the composer I’ll link to below.

✦𝑯𝑶𝑹𝑺𝑴𝑨✦

32,413 views • 4 months ago

Today, we’re excited to announce Deep Extract. We’re pushing the boundary of complex document extraction by utilizing an agent harness approach, iterating and verifying outputs until they are at human-level accuracy. The results speak for themselves: we’ve handled documents exceeding 2,500 pages, extracting a total of 30M+ fields in the last couple weeks with results in our production beta reaching 99-100% accuracy. It’s particularly effective on complex documents with long lists, such as invoice line items, brokerage statement transactions, equipment manifests, and more. See more on how we built it, how to use it, and more👇

Today, we’re excited to announce Deep Extract. We’re pushing the boundary of complex document extraction by utilizing an agent harness approach, iterating and verifying outputs until they are at human-level accuracy. The results speak for themselves: we’ve handled documents exceeding 2,500 pages, extracting a total of 30M+ fields in the last couple weeks with results in our production beta reaching 99-100% accuracy. It’s particularly effective on complex documents with long lists, such as invoice line items, brokerage statement transactions, equipment manifests, and more. See more on how we built it, how to use it, and more👇

Reducto

23,949,618 views • 6 months ago

Over 1 billion PDFs are created every day, but your agents still can’t read them reliably. Today we’re releasing Parse 2.0, the most accurate document parsing API in the world. Extend already processes millions of pages daily for leading AI teams like Brex, Mercury, Opendoor, Flatiron Health, and hundreds of others. Now, its even better. Parse 2.0 is SOTA quality on RealDoc-Bench, our open source benchmark that measures agent success rate on real world docs that agents actually encounter in production. We trained Parse 2.0 on 1M+ pages of the hardest documents seen in production. Here’s how it stacks up: - #1 in healthcare, real estate, logistics, and financial services - 95.7% agent Q&A accuracy on 581 docs (next best: 92%) - 0.847 F1 on layout (next best: 0.759) Give it a try today and build production-ready document agents with Extend.

Over 1 billion PDFs are created every day, but your agents still can’t read them reliably. Today we’re releasing Parse 2.0, the most accurate document parsing API in the world. Extend already processes millions of pages daily for leading AI teams like Brex, Mercury, Opendoor, Flatiron Health, and hundreds of others. Now, its even better. Parse 2.0 is SOTA quality on RealDoc-Bench, our open source benchmark that measures agent success rate on real world docs that agents actually encounter in production. We trained Parse 2.0 on 1M+ pages of the hardest documents seen in production. Here’s how it stacks up: - #1 in healthcare, real estate, logistics, and financial services - 95.7% agent Q&A accuracy on 581 docs (next best: 92%) - 0.847 F1 on layout (next best: 0.759) Give it a try today and build production-ready document agents with Extend.

Kushal Byatnal

589,622 views • 5 months ago

Been wondering how to use Decision Models in your agent harness? Introducing Sage Router, in early access on Antseed: - 39% cheaper than Opus at the same accuracy, - and 31% more efficient than JevRouter - On OpenRouter's leaderboard (GPQA Diamond), it matches the top router's accuracy at a third of the cost. Benchmarks👇 Now live on the Antseed Network.

Been wondering how to use Decision Models in your agent harness? Introducing Sage Router, in early access on Antseed: - 39% cheaper than Opus at the same accuracy, - and 31% more efficient than JevRouter - On OpenRouter's leaderboard (GPQA Diamond), it matches the top router's accuracy at a third of the cost. Benchmarks👇 Now live on the Antseed Network.

Levanto Labs

33,999 views • 7 days ago

We built an AI agent that lets you vibe-code document extraction - high accuracy and citations over the most complex documents. Our latest release lets you upload documents as context. All you then have to do is describe what you want extracted in natural language. 💡 Our agent will then read the document with file tools to infer the right schema, validation rules, and other pre/postprocessing logic. ✅ It will give you back a workflow that can extract over thousands/millions of documents at scale. You can still of course review and edit every output before approving. Stop handling paperwork manually; just upload files, describe your task, and let our agent handle the rest. Our vision for LlamaAgents is to provide the most advanced and easy-to-use way for you to orchestrate document work. Walkthrough: Check it out: If you’re interested in reducing the operational burden of document extraction (invoices, claims, onboarding forms), come talk to us!

We built an AI agent that lets you vibe-code document extraction - high accuracy and citations over the most complex documents. Our latest release lets you upload documents as context. All you then have to do is describe what you want extracted in natural language. 💡 Our agent will then read the document with file tools to infer the right schema, validation rules, and other pre/postprocessing logic. ✅ It will give you back a workflow that can extract over thousands/millions of documents at scale. You can still of course review and edit every output before approving. Stop handling paperwork manually; just upload files, describe your task, and let our agent handle the rest. Our vision for LlamaAgents is to provide the most advanced and easy-to-use way for you to orchestrate document work. Walkthrough: Check it out: If you’re interested in reducing the operational burden of document extraction (invoices, claims, onboarding forms), come talk to us!

Jerry Liu

20,857 views • 8 months ago

Announcing Agent Composer: AI that works when it actually IS rocket science. Technical teams can now automate the routine (but complex) tasks that used to take hours every week—root cause analysis, production planning, test code generation— and reduce them down to minutes, so you can get your real work done. Here's how we solved the AI context problem for hard engineering tasks (like rocket science 🚀) 🧵

Announcing Agent Composer: AI that works when it actually IS rocket science. Technical teams can now automate the routine (but complex) tasks that used to take hours every week—root cause analysis, production planning, test code generation— and reduce them down to minutes, so you can get your real work done. Here's how we solved the AI context problem for hard engineering tasks (like rocket science 🚀) 🧵

Contextual AI

1,553,121 views • 9 months ago

Open Design 0.8.0 now works inside Cursor Composer 2.5. Composer 2.5 is built for this: long tasks, tool use, and iteration in one run. So we asked Composer to control Open Design through chat and build an animated HTML website. What happened 2.5 minutes later is insane: - landing page - smooth motion via GSAP gsap-skills - responsive HTML - ready to iterate Here is a quick demo:

Open Design 0.8.0 now works inside Cursor Composer 2.5. Composer 2.5 is built for this: long tasks, tool use, and iteration in one run. So we asked Composer to control Open Design through chat and build an animated HTML website. What happened 2.5 minutes later is insane: - landing page - smooth motion via GSAP gsap-skills - responsive HTML - ready to iterate Here is a quick demo:

Open Design

16,926 views • 5 months ago

TestSprite 3.0 is live. A fleet of parallel AI agents that use your app like real users — in minutes. This is the release we've been building all year. The numbers tell the story of our completely overhauled engine: - Accuracy jumped by nearly 40% on the hardest, most complex projects. - Coverage increased significantly across the board for every E2E run. - Reliability is locked in—stable and consistent across multiple runs. We rebuilt our entire engine around these performance gains so you can ship fast and ship confident.

TestSprite 3.0 is live. A fleet of parallel AI agents that use your app like real users — in minutes. This is the release we've been building all year. The numbers tell the story of our completely overhauled engine: - Accuracy jumped by nearly 40% on the hardest, most complex projects. - Coverage increased significantly across the board for every E2E run. - Reliability is locked in—stable and consistent across multiple runs. We rebuilt our entire engine around these performance gains so you can ship fast and ship confident.

TestSprite

814,741 views • 5 months ago

With the Cursor SDK, you can build your own agents with Composer 2.5. It's now available in Python and TypeScript. This long weekend, Composer usage is 90% off in the SDK. We're excited to see what you build!

With the Cursor SDK, you can build your own agents with Composer 2.5. It's now available in Python and TypeScript. This long weekend, Composer usage is 90% off in the SDK. We're excited to see what you build!

Cursor

593,582 views • 5 months ago

Introducing Parrot 🦜 Our first speech-to-text model built for production-grade voice agents with an industry leading accuracy, built for noisy calls, regional dialects and compressed audio. Trained on thousands of hours of real-world audio to make sure when you say Udaipur, the agent doesn't hear 'you-die-poor'.

Introducing Parrot 🦜 Our first speech-to-text model built for production-grade voice agents with an industry leading accuracy, built for noisy calls, regional dialects and compressed audio. Trained on thousands of hours of real-world audio to make sure when you say Udaipur, the agent doesn't hear 'you-die-poor'.

Vedant

15,977 views • 5 months ago

Audio Transcription with Google Gemini 1.5 Flash In this video, Gemini-Flash was able to transcribe 13 minutes of audio in 50-60 seconds. I have tested this with multiple audio files, and the transcription accuracy is close to 99%. If the audio is clear, you get 100% correct transcription. Even if the audio is really bad with a lot of noise, you still get around 95-96% accuracy. The code for this is available on GitHub. If you're interested, you can download and run it. You just need a Google API key.

Audio Transcription with Google Gemini 1.5 Flash In this video, Gemini-Flash was able to transcribe 13 minutes of audio in 50-60 seconds. I have tested this with multiple audio files, and the transcription accuracy is close to 99%. If the audio is clear, you get 100% correct transcription. Even if the audio is really bad with a lot of noise, you still get around 95-96% accuracy. The code for this is available on GitHub. If you're interested, you can download and run it. You just need a Google API key.

AshutoshShrivastava

123,061 views • 2 years ago

Heartwarming: This Israeli-American composer gives a live piano performance in an airport for passerby on their way to the Caribbean Repost if you have been touched by him

Heartwarming: This Israeli-American composer gives a live piano performance in an airport for passerby on their way to the Caribbean Repost if you have been touched by him

White Boy Ganyu ☦️

244,741 views • 5 months ago

How to use the new Icon Composer with an existing Xcode project: 1. Make a new icon file 2. Drag it into your Xcode project (Move into) 3. Xcode should prioritize that over previous AppIcon files 4. Close the Icon Composer file 5. Click on Icon in Xcode 6. Open in Icon Composer (Make edits now) Things to Note: 1. Xcode will render the new static icon for previous macOS versions (Sequioa) 2. If you drag the app icon around your Dock, you may not see the app icon update unless you drag it off the Dock (Stop the app) 3. You may need to Clean + Run to see the icon update 4. Change the "Primary App Icon Set Name" to choose between multiple Composer files.

How to use the new Icon Composer with an existing Xcode project: 1. Make a new icon file 2. Drag it into your Xcode project (Move into) 3. Xcode should prioritize that over previous AppIcon files 4. Close the Icon Composer file 5. Click on Icon in Xcode 6. Open in Icon Composer (Make edits now) Things to Note: 1. Xcode will render the new static icon for previous macOS versions (Sequioa) 2. If you drag the app icon around your Dock, you may not see the app icon update unless you drag it off the Dock (Stop the app) 3. You may need to Clean + Run to see the icon update 4. Change the "Primary App Icon Set Name" to choose between multiple Composer files.

Paul Solt

15,480 views • 6 months ago

We tuned an AI agent that can do large-scale document extraction from long docs (50+ pages, some with 10k-100k fields) with 94%+ accuracy 📈 It uses a harness + model set that is tuned specifically for reasoning over extracting out complex information from complex docs. Each extracted field comes with a confidence score as well as a bounding box denoting where it came from. It does 10-20% better in accuracy than generalized coding agent harnesses (e.g. Claude Code Opus 4.8 and Codex GPT-5.6). Check out the video below for a demo. The mode is called LlamaExtract Agentic Plus. Learn more about our extraction benchmark and agentic plus mode here: If you have very complex document extraction needs, come check out LlamaParse!

We tuned an AI agent that can do large-scale document extraction from long docs (50+ pages, some with 10k-100k fields) with 94%+ accuracy 📈 It uses a harness + model set that is tuned specifically for reasoning over extracting out complex information from complex docs. Each extracted field comes with a confidence score as well as a bounding box denoting where it came from. It does 10-20% better in accuracy than generalized coding agent harnesses (e.g. Claude Code Opus 4.8 and Codex GPT-5.6). Check out the video below for a demo. The mode is called LlamaExtract Agentic Plus. Learn more about our extraction benchmark and agentic plus mode here: If you have very complex document extraction needs, come check out LlamaParse!

Jerry Liu

18,359 views • 2 months ago

Josh Allen’s accuracy is borderline ROBOTIC at this point. Allen reportedly hit 7 nets IN A ROW at practice today. As a prospect, multiple NFL scouts tabbed Allen’s accuracy as his “biggest” red flag. Even as a rookie, Allen was one of least accurate passers in the league. Now, he’s arguably Top-5 in accuracy amongst all QB’s in the league. The Bills put on an absolute CLINIC in QB development with Allen 😳

Josh Allen’s accuracy is borderline ROBOTIC at this point. Allen reportedly hit 7 nets IN A ROW at practice today. As a prospect, multiple NFL scouts tabbed Allen’s accuracy as his “biggest” red flag. Even as a rookie, Allen was one of least accurate passers in the league. Now, he’s arguably Top-5 in accuracy amongst all QB’s in the league. The Bills put on an absolute CLINIC in QB development with Allen 😳

Rookie Watch

352,203 views • 1 year ago

Composer: Okay, so what did you have in mind for the back-up sound? Honda: I want an SATB chorus singing a chord that will please the lord. Composer: you got it.

Composer: Okay, so what did you have in mind for the back-up sound? Honda: I want an SATB chorus singing a chord that will please the lord. Composer: you got it.

score follower

34,885 views • 2 years ago

@MarkieWagner explaining to @TBPN how Poetic is able to run multi-hour complex processes for F500s with 99% accuracy in production, all with minimal token spend.

@MarkieWagner explaining to @TBPN how Poetic is able to run multi-hour complex processes for F500s with 99% accuracy in production, all with minimal token spend.

Poetic

23,339 views • 4 months ago

Did you know that the composer of the famous Game of Thrones soundtrack is Iranian? You are now listening to a cover of this iconic piece on the Iranian Daf, arranged by Reza Sajjadi and the original composer of this music is Ramin Djawadi. ENJOY!

Did you know that the composer of the famous Game of Thrones soundtrack is Iranian? You are now listening to a cover of this iconic piece on the Iranian Daf, arranged by Reza Sajjadi and the original composer of this music is Ramin Djawadi. ENJOY!

Iran Military Media ☫

158,553 views • 8 months ago