Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Built a vibe-coded presentation app generator that turns natural language into polished slides ✨ This project by Jerry Liu combines the Claude Agent SDK with LlamaParse to create an AI-powered presentation tool that handles everything from content creation to PDF export: 🎯 Chat-based slide creation - just describe what... you want and watch slides appear ✏️ Real-time editing through natural conversation - refine slides by chatting with the AI 📄 Smart document parsing with LlamaParse for incorporating reference materials 📊 Full export functionality to PowerPoint and PDF formats Perfect example of how LlamaParse makes document processing seamless while Claude's conversational abilities create an intuitive slide editing experience. Check out the full project:show more

LlamaIndex 🦙

106,708 subscribers

50,184 просмотров • 8 месяцев назад •via X (Twitter)

Наука и технологии

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

We’re excited to officially launch LlamaParse, the first genAI-native document parsing solution. Not only is it better at parsing out images/tables/charts 📊📈 than virtually every other parser, it is now steerable through natural language instructions - output the document in whatever format you desire! It is also the only parsing solution that seamlessly allows you to build accurate RAG over complex documents, free of hallucinations 🔥 We launched it in private preview a few weeks ago and hit 2k users, 1M total PDF pages parsed. And now it’s better than ever. LlamaParse contains the following killer features: ✅ SOTA table/chart extraction ✅ Seamless integration with LlamaIndex 🦙 advanced RAG/agents ✅✨ Natural language Parsing Instructions ✅✨JSON mode and image extraction ✅✨Support for ~10 document types (.pdf, .pptx, .docx, .xml) and more Our pricing is simple: 1k free per day, and additional pages at 0.3c a page, or $3 for 1k pages. If you want advanced document RAG and/or private deployments, come get in touch with us to chat about LlamaCloud. Check out our full blog post here: LlamaParse client repo: Signup at 🦙☁️: Come talk to us:

We’re excited to officially launch LlamaParse, the first genAI-native document parsing solution. Not only is it better at parsing out images/tables/charts 📊📈 than virtually every other parser, it is now steerable through natural language instructions - output the document in whatever format you desire! It is also the only parsing solution that seamlessly allows you to build accurate RAG over complex documents, free of hallucinations 🔥 We launched it in private preview a few weeks ago and hit 2k users, 1M total PDF pages parsed. And now it’s better than ever. LlamaParse contains the following killer features: ✅ SOTA table/chart extraction ✅ Seamless integration with LlamaIndex 🦙 advanced RAG/agents ✅✨ Natural language Parsing Instructions ✅✨JSON mode and image extraction ✅✨Support for ~10 document types (.pdf, .pptx, .docx, .xml) and more Our pricing is simple: 1k free per day, and additional pages at 0.3c a page, or $3 for 1k pages. If you want advanced document RAG and/or private deployments, come get in touch with us to chat about LlamaCloud. Check out our full blog post here: LlamaParse client repo: Signup at 🦙☁️: Come talk to us:

LlamaIndex 🦙

143,198 просмотров • 2 лет назад

REPLIT JUST LAUNCHED AI SLIDES THAT DON'T LOOK LIKE AI MADE THEM every other AI slide generator spits out decks that scream "made by AI." replit slides is going hard on the angle that yours will actually look presentable. how it works: > describe what you want and it builds a full deck > start from scratch or use a built-in template > import an existing pptx and let AI restyle the whole thing > export to pptx, pdf, or google slides the actually interesting part: > slides are generated as code under the hood > same project can also spin up apps, mobile UIs, dashboards, and animations > everything shares the same design tokens (fonts, colors, spacing) > so your launch deck and the live dashboard it links to look like one product this isn't really replit competing with gamma or tome. its replit betting that slides, dashboards, animations, and apps are all just code (and once you have an agent that writes code well, every knowledge artifact comes from the same prompt). link in replies 👇

REPLIT JUST LAUNCHED AI SLIDES THAT DON'T LOOK LIKE AI MADE THEM every other AI slide generator spits out decks that scream "made by AI." replit slides is going hard on the angle that yours will actually look presentable. how it works: > describe what you want and it builds a full deck > start from scratch or use a built-in template > import an existing pptx and let AI restyle the whole thing > export to pptx, pdf, or google slides the actually interesting part: > slides are generated as code under the hood > same project can also spin up apps, mobile UIs, dashboards, and animations > everything shares the same design tokens (fonts, colors, spacing) > so your launch deck and the live dashboard it links to look like one product this isn't really replit competing with gamma or tome. its replit betting that slides, dashboards, animations, and apps are all just code (and once you have an agent that writes code well, every knowledge artifact comes from the same prompt). link in replies 👇

Creators Toolbox

29,232 просмотров • 5 месяцев назад

Check out the form-filling agent that automates PDF forms using AI by Jerry Liu 📄🤖 Use any fillable PDF with an agent that fills it out based on your prompts and context files. Our new experiment creates a multi-turn chat experience for form completion. 🔍 Upload fillable PDFs and automatically detect form fields using PyMuPDF 📝 Add custom prompts and context files (parsed via LlamaParse) to guide the AI 🤖 Multi-turn conversations let you refine and correct form entries after initial completion 💾 Download your completed forms when done The agent uses simple tools to list, set, get, and validate form fields. You can chat with it to make corrections and adjustments until your form is perfect. Check out the code on GitHub: Or the deployed app here:

Check out the form-filling agent that automates PDF forms using AI by Jerry Liu 📄🤖 Use any fillable PDF with an agent that fills it out based on your prompts and context files. Our new experiment creates a multi-turn chat experience for form completion. 🔍 Upload fillable PDFs and automatically detect form fields using PyMuPDF 📝 Add custom prompts and context files (parsed via LlamaParse) to guide the AI 🤖 Multi-turn conversations let you refine and correct form entries after initial completion 💾 Download your completed forms when done The agent uses simple tools to list, set, get, and validate form fields. You can chat with it to make corrections and adjustments until your form is perfect. Check out the code on GitHub: Or the deployed app here:

LlamaIndex 🦙

55,131 просмотров • 9 месяцев назад

LiteParse Samples 📄🧑‍🏫 LiteParse is our free and fast document parser that can parse text with bounding boxes from any document. I’ve created a new repository that contains some demos on how you can make use of its outputs! ✅ Comparison against PyPDF and PyMuPDF ✅ Visual Citations: search for any keyword and see joined bounding boxes pop up in the source images It should be a default tool to help AI agents OCR any document type (supports PDF, Word, Powerpoint + dozens of other formats), before heavier-weight VLM-based document parsing solutions (like our own LlamaParse). LiteParse Samples repo: LiteParse:

LiteParse Samples 📄🧑‍🏫 LiteParse is our free and fast document parser that can parse text with bounding boxes from any document. I’ve created a new repository that contains some demos on how you can make use of its outputs! ✅ Comparison against PyPDF and PyMuPDF ✅ Visual Citations: search for any keyword and see joined bounding boxes pop up in the source images It should be a default tool to help AI agents OCR any document type (supports PDF, Word, Powerpoint + dozens of other formats), before heavier-weight VLM-based document parsing solutions (like our own LlamaParse). LiteParse Samples repo: LiteParse:

Jerry Liu

13,196 просмотров • 6 месяцев назад

How to turn a PowerPoint into a full video with voiceover (without recording anything). Turning slides into a video meant: ↳ Editing timelines, recording audio, fixing pacing, basically rebuilding your deck from scratch. Not anymore… With Synthesia 🎥 you just upload a PPT and get a video: ➟ Your slides become editable video scenes ➟ Your speaker notes become a natural voiceover ➟ An AI avatar delivers your content on-screen ➟ Your videos can be translated into 140+ languages ➟ All while keeping your original layout, fonts, and design intact No camera setup. No retakes. No audio editing. ↳ And this feature is free to use. Just: upload a PPT → AI generate video → download it. This turns your "static" assets into high-value content: ✓ Training deck sitting in your Google Drive? Now a watchable video. ✓ Sales slides nobody reads? Now a 3-minute demo with a presenter. ✓ Onboarding presentation? Now a proper video course. 👉 And the best part: It’s a free feature. 📌 Convert PowerPoints to videos for free: 🔄 Repost to help others turn slides into videos — #AI #Video #PowerPoint #SynthesiaPartnership

Andrew Bolis

41,775 просмотров • 9 месяцев назад

Introducing Genspark AI Slides 5.0. Our biggest upgrade to AI Slides yet. The biggest leap? Slide Skills. Reusable assets that bundle expert thinking and polished design, ready in one click. Here's everything new in 5.0: - 100+ Expert Skills built in. We've packed the thinking of the world's top experts, across consulting, sales, teaching, and more. Apply with one click, and their expertise becomes yours. - Build your own Skill. Upload a deck or PDF, describe the analysis steps, and lock in your team's style. - 5× faster, higher quality. Powered by next-gen architecture and the most powerful models. - Built-in data analysis. Crunches your Excel and Sheets data into presentation-ready insights. Every deck thinks like an expert. Try it now → 🎉 50% off AI Slides 5.0 credits, 14 days only

Introducing Genspark AI Slides 5.0. Our biggest upgrade to AI Slides yet. The biggest leap? Slide Skills. Reusable assets that bundle expert thinking and polished design, ready in one click. Here's everything new in 5.0: - 100+ Expert Skills built in. We've packed the thinking of the world's top experts, across consulting, sales, teaching, and more. Apply with one click, and their expertise becomes yours. - Build your own Skill. Upload a deck or PDF, describe the analysis steps, and lock in your team's style. - 5× faster, higher quality. Powered by next-gen architecture and the most powerful models. - Built-in data analysis. Crunches your Excel and Sheets data into presentation-ready insights. Every deck thinks like an expert. Try it now → 🎉 50% off AI Slides 5.0 credits, 14 days only

Genspark

210,765 просмотров • 4 месяцев назад

Last week, Anthropic dropped the coolest "AI isn't just chat" product. Claude Design lets you describe what you want to Claude and it returns prototypes, slides, and one-pagers by just chatting. You can then export to Canva, PDF, PPT, or hand off to Claude Code. You can give your entire design system and codebase to apply automatically to your project, and can share and collaborate with your team. The direct connect to GitHub will probably be my most used feature. Feels directly like a competitor to Figma or just a way to further boost collaboration with Claude. I built a preview of an app that allows CSGO players to meet and squad up with other players, marketplace for in-game loadouts, and pay high-ranking players to mentor them. (CSGO is a game released in 2012 and still consistently one of the highest watched games on Twitch, extremely popular in esports, highly competitive, and an all-around classic game - I know this because a gen zer talked to me about it for 2 straight hours.) I'll keep sharing more examples here.

Last week, Anthropic dropped the coolest "AI isn't just chat" product. Claude Design lets you describe what you want to Claude and it returns prototypes, slides, and one-pagers by just chatting. You can then export to Canva, PDF, PPT, or hand off to Claude Code. You can give your entire design system and codebase to apply automatically to your project, and can share and collaborate with your team. The direct connect to GitHub will probably be my most used feature. Feels directly like a competitor to Figma or just a way to further boost collaboration with Claude. I built a preview of an app that allows CSGO players to meet and squad up with other players, marketplace for in-game loadouts, and pay high-ranking players to mentor them. (CSGO is a game released in 2012 and still consistently one of the highest watched games on Twitch, extremely popular in esports, highly competitive, and an all-around classic game - I know this because a gen zer talked to me about it for 2 straight hours.) I'll keep sharing more examples here.

Allie K. Miller

10,586 просмотров • 6 месяцев назад

Turning a simple product idea into real-world packaging has never been this seamless! I tested out an AI-powered design workflow with Packify AI to create a high-end luxury look for a skincare concept from scratch. From chatting with AI to build the layout, editing layers, and rotating the 3D preview, everything happens in one place without touching complex design software. And the best part? You can instantly export factory-ready dielines for physical production in just one click. If you are an entrepreneur, creator, or running an e-commerce brand, you need to check this out. Try it free at and use code SIMPLY20 for 20% off! #AIDesign #PackifyDesign #PackifyAI #PackagingDesign #AITools

Anissa

14,037 просмотров • 2 месяцев назад

Most AI video tools help you edit. Vizard Agent helps you recreate an entire creative workflow. The Reference Video feature is what stood out to me. Simply give Vizard Agent a viral video, ad, or any piece of content as a reference. It analyzes the storytelling, pacing, editing style, shot sequence, and overall creative direction. Then it tells you exactly what footage you need, identifies any missing assets, and guides you through the production process. Once you upload your footage, the Agent takes over. It selects the best moments, structures the story, adds captions and visuals, makes creative editing decisions, and delivers a polished video inspired by the reference while keeping it original. For creators, brands, and agencies, this opens up an entirely new workflow. This is more than AI editing. It's delegating the entire video project to an AI that understands your goal and executes it from brief to final export. That's what makes Vizard Agent feel like one of the biggest steps toward true Video AGI.

Most AI video tools help you edit. Vizard Agent helps you recreate an entire creative workflow. The Reference Video feature is what stood out to me. Simply give Vizard Agent a viral video, ad, or any piece of content as a reference. It analyzes the storytelling, pacing, editing style, shot sequence, and overall creative direction. Then it tells you exactly what footage you need, identifies any missing assets, and guides you through the production process. Once you upload your footage, the Agent takes over. It selects the best moments, structures the story, adds captions and visuals, makes creative editing decisions, and delivers a polished video inspired by the reference while keeping it original. For creators, brands, and agencies, this opens up an entirely new workflow. This is more than AI editing. It's delegating the entire video project to an AI that understands your goal and executes it from brief to final export. That's what makes Vizard Agent feel like one of the biggest steps toward true Video AGI.

Leonardo

91,608 просмотров • 2 месяцев назад

I solved building decks with AI agents — by giving them a CLI tool like Powerpoint or Google Slides. AI could already make a beautiful deck if you asked it to using Ant's pptx skill. The problem was working with it. If it made one alignment mistake, fixing it on one slide would break something on another, and it became a game of whack-a-mole. One time I spent two days playing AI roulette, hoping the next prompt would finally fix the thing, and ended up building the whole deck by hand because I was on a deadline. So I built Hands-on Deck. And the reason it works is that this isn't just a skill — this is PowerPoint. The actual application: PowerPoint, Google Slides, Keynote, whatever you use. This is that, but for an agent, presented as a CLI. Every gesture you make in a deck app maps to a command. Click a box and type, drag a shape from here to there, look at a slide – agent can do it all in a command. And that changes how the agent behaves. With this CLI it works and thinks like a designer — it looks, makes an edit, looks again, makes another surgical edit. Compare that to Anthropic's pptx skill, built on the idea that Claude is a great programmer: it literally writes code to manipulate the deck, hand-editing XML and hoping it doesn't break anything else in the middle. The real test isn't creating something once — it's whether it can make surgical edits like you want. That's what I did in this video walkthrough and my claude crushed it! Check it out for yourself. So decks can be built like a designer now — with real flavor and taste. If you spend hours every week on decks, this gives those hours back. You can install it as a skill in Claude Code, Codex, whatever you use. Works every harness that supports skills. Let me know if you make something cool with it.

I solved building decks with AI agents — by giving them a CLI tool like Powerpoint or Google Slides. AI could already make a beautiful deck if you asked it to using Ant's pptx skill. The problem was working with it. If it made one alignment mistake, fixing it on one slide would break something on another, and it became a game of whack-a-mole. One time I spent two days playing AI roulette, hoping the next prompt would finally fix the thing, and ended up building the whole deck by hand because I was on a deadline. So I built Hands-on Deck. And the reason it works is that this isn't just a skill — this is PowerPoint. The actual application: PowerPoint, Google Slides, Keynote, whatever you use. This is that, but for an agent, presented as a CLI. Every gesture you make in a deck app maps to a command. Click a box and type, drag a shape from here to there, look at a slide – agent can do it all in a command. And that changes how the agent behaves. With this CLI it works and thinks like a designer — it looks, makes an edit, looks again, makes another surgical edit. Compare that to Anthropic's pptx skill, built on the idea that Claude is a great programmer: it literally writes code to manipulate the deck, hand-editing XML and hoping it doesn't break anything else in the middle. The real test isn't creating something once — it's whether it can make surgical edits like you want. That's what I did in this video walkthrough and my claude crushed it! Check it out for yourself. So decks can be built like a designer now — with real flavor and taste. If you spend hours every week on decks, this gives those hours back. You can install it as a skill in Claude Code, Codex, whatever you use. Works every harness that supports skills. Let me know if you make something cool with it.

Nityesh

70,086 просмотров • 3 месяцев назад

Video creation just got a lot more interesting. (You might not need a complicated editing workflow to create professional videos anymore.) Synthesia 🎥 just introduced Syren, and the idea behind it is pretty impressive. You describe the video you want, and Syren creates the entire thing for you. Here’s what makes it different. → It learns your brand style from existing videos. → It creates motion graphics, AI avatars, voiceovers, music, and B-roll. → You can start with a simple prompt, document, or link. → You can refine your video through a chat interface. But the most interesting part? Syren learns how your brand looks and moves, from typography and transitions to pacing and editing style. That means you can create new videos that stay consistent with your brand, without starting from scratch every time. Think product launches, feature announcements, explainers, and tutorials. No complicated editing workflow. Just your idea and a much simpler way to bring it to life. This isn't about replacing designers or editors. It's about making professional video creation accessible to more people. And I think that's a pretty big shift. Want to try it yourself? Sign up to get 2,000 free video credits. #Synthesia #AI #ad

Darshal Jaitwar

23,814 просмотров • 11 дней назад

You have an idea for a cinematic Tesla Roadster ad. You know the vibe, the visual direction, and the story you want to tell. Now, how do you turn it into a complete video without getting lost in the process? That's what I wanted to explore with CapCut Web × Codex. By connecting CapCut through the plugin, I can describe my creative goals in natural language and let Codex call CapCut to carry out the video creation workflow. No need to start by mapping out every technical operation. Start with the idea and let the tools help move it forward. It's a more conversational way to approach AI video editing, bringing creative intent and execution closer together. #capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing

You have an idea for a cinematic Tesla Roadster ad. You know the vibe, the visual direction, and the story you want to tell. Now, how do you turn it into a complete video without getting lost in the process? That's what I wanted to explore with CapCut Web × Codex. By connecting CapCut through the plugin, I can describe my creative goals in natural language and let Codex call CapCut to carry out the video creation workflow. No need to start by mapping out every technical operation. Start with the idea and let the tools help move it forward. It's a more conversational way to approach AI video editing, bringing creative intent and execution closer together. #capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing

Jay

51,982 просмотров • 1 день назад

Open-sourcing GenOffice, the world's first full-featured open-source AI Office for PC and Mac. GenOffice is free for everyone, ad-free. All the editing tools you'd expect, no strings attached. How it started? One engineer, one week, $10,000 in tokens. That's what it took to build the GenOffice Alpha. Help shape the AI office experience We invite you to build the rest of it with us. Join the GenOffice group chat on GenTeam and bring your requests straight to the agent. Not by writing code, but by telling it what you want. You say what you need, and every piece of your feedback becomes part of the product. What's in GenOffice? Docs, Sheets, Slides, and PDF, with all the editing tools you already know. And with Genspark Super Agent built in, powered by top-tier models, it digs into research, crunches your data, then writes the doc or builds the deck for you, using your Genspark credits. This is an Alpha, and we mean it. Come tell us what's broken and what's missing. Everyone who jumps in with real feedback gets 1,000+ Genspark credits as a thank you. GitHub 👉 Download it now 👉

Open-sourcing GenOffice, the world's first full-featured open-source AI Office for PC and Mac. GenOffice is free for everyone, ad-free. All the editing tools you'd expect, no strings attached. How it started? One engineer, one week, $10,000 in tokens. That's what it took to build the GenOffice Alpha. Help shape the AI office experience We invite you to build the rest of it with us. Join the GenOffice group chat on GenTeam and bring your requests straight to the agent. Not by writing code, but by telling it what you want. You say what you need, and every piece of your feedback becomes part of the product. What's in GenOffice? Docs, Sheets, Slides, and PDF, with all the editing tools you already know. And with Genspark Super Agent built in, powered by top-tier models, it digs into research, crunches your data, then writes the doc or builds the deck for you, using your Genspark credits. This is an Alpha, and we mean it. Come tell us what's broken and what's missing. Everyone who jumps in with real feedback gets 1,000+ Genspark credits as a thank you. GitHub 👉 Download it now 👉

Genspark

519,756 просмотров • 2 месяцев назад

I just built a branded IG carousel generator in Claude Code 🤯 One brand URL + one product name = 6 finished carousel slides. The kind an agency charges $2K to produce. All inside Claude Code. Perfect for DTC brands and agencies who need to post carousels every week but can't turn each one into a two-day design project. If every carousel is the same mini-project — writing the copy in Notion, rebuilding the template in Canva, exporting slides one at a time, then three rounds of revisions with your designer before it's even live... This tool eliminates the entire loop: → Drop in a brand URL and product name → Claude scrapes the site and extracts colors, fonts, voice, and positioning → Generates 6 slide concepts across proven carousel frameworks → Writes a detailed image prompt for each slide → Fires all 6 to ChatGPT Images 2.0 via FAL in parallel → Downloads finished slides + opens an HTML gallery No Canva. No designer back-and-forth. No generic AI slop. What you get: → Finished 1080x1350px slides ready to post directly to Instagram → Wavy color-blocked backgrounds, bold typography, product hero — all baked into the image → Brand-accurate colors and copy pulled from the live site automatically → A reusable pipeline — new brand, new folder, same 3-minute workflow Built 100% in Claude Code. I put together the full step-by-step playbook so you can build this yourself. Want it for free? Like this post Comment "SLIDES" And I'll send it over (must be following so I can DM)

Mike Futia

36,550 просмотров • 3 месяцев назад

i just cracked the code on ai digital products... every product costs $0 to build and takes under 2 minutes to make. i built an ai agent that turns any niche into a ready-to-export digital product in whatever format you want, pdf, notion template, ebook, guide, you name it. drop in your niche and one problem. the system analyzes demand, maps buyer desires, builds you the full product, and you watch it happen in real time. then the agent gives you the path, what to do next, rates each module, edits everything, practically sells it for you. you just follow it. i'm making full end-to-end products in under 2 minutes with this system. so i packaged the whole thing up. you're getting an ai that: - picks your niche - calculates demand - builds the product in any format - gives you the selling blueprint to go with it. rt + reply "product" and i'll show you how to access (must be following so i can dm)

i just cracked the code on ai digital products... every product costs $0 to build and takes under 2 minutes to make. i built an ai agent that turns any niche into a ready-to-export digital product in whatever format you want, pdf, notion template, ebook, guide, you name it. drop in your niche and one problem. the system analyzes demand, maps buyer desires, builds you the full product, and you watch it happen in real time. then the agent gives you the path, what to do next, rates each module, edits everything, practically sells it for you. you just follow it. i'm making full end-to-end products in under 2 minutes with this system. so i packaged the whole thing up. you're getting an ai that: - picks your niche - calculates demand - builds the product in any format - gives you the selling blueprint to go with it. rt + reply "product" and i'll show you how to access (must be following so i can dm)

MONTE

16,596 просмотров • 3 месяцев назад

The latest RAG trend for the current agent harnesses (Codex, Cowork) is to do two passes of document processing to solve a knowledge work task over a data room of documents: 1️⃣ A fast and light pass, oftentimes using a free/OSS doc parsing tool. This can be cheaply run across 10-100-1k’s of files, and enables the agent to then do retrieval (e.g. grep, semantic) to find relevant subsets of context. 2️⃣ A “just-in-time” VLM-based pass. Once the agent finds the relevant pages of context, it will screenshot the documents can call its own VLM (or write code) to dissect the pages. The issue with only using VLM-based OCR tools over massive ad-hoc customer file dumps is that it’s slow and expensive. Doing JIT VLM OCR allows the agent to filter through the data cheaply, but still preserve accuracy for the context that’s needed for the task. The agent harnesses do two-pass document processing by default using off the shelf-tools: pdf2text as the first pass, and using itself (Opus 5) as the second pass. See the below video where Cowork runs over a bunch of PDFs to answer a question about a benchmark graph in the Kimi k3 paper. The main issues here with the “out of the box” doc processing these agents offer are: * Opus 5 is not the best VLM for OCR. It is also way too expensive at scale and lacks grounding * The OSS tools like pypdf, pdf2text, may not be versatile enough as the first pass. * The agent will write a lot of throwaway code to rewrite things an OCR tool would’ve provided out of the box, like chart processing, bounding boxes, confidence scores, leading to increased cost and speed. We have all the tools within LlamaIndex 🦙 to help any agent do two-pass document processing with higher accuracy and lower cost. 1️⃣ We have liteparse for the first pass - a free/OSS parser written in Rust that’s faster/more accurate than other OSS parsers, and supports 50+ document types 2️⃣ We have LlamaParse for the second pass - an agentic document engine that uses VLMs+harnesses to achieve SOTA in accuracy and cost across various doc parsing and extraction tasks. It can be called from any agent harness as an MCP or skill. It takes in page numbers as input, so that the agent can choose to run LlamaParse over a subset of the doc instead of the full doc as a “zoom-in” pass. Come check it out! LiteParse: LlamaParse: All the relevant docs, including MCP, are here:

The latest RAG trend for the current agent harnesses (Codex, Cowork) is to do two passes of document processing to solve a knowledge work task over a data room of documents: 1️⃣ A fast and light pass, oftentimes using a free/OSS doc parsing tool. This can be cheaply run across 10-100-1k’s of files, and enables the agent to then do retrieval (e.g. grep, semantic) to find relevant subsets of context. 2️⃣ A “just-in-time” VLM-based pass. Once the agent finds the relevant pages of context, it will screenshot the documents can call its own VLM (or write code) to dissect the pages. The issue with only using VLM-based OCR tools over massive ad-hoc customer file dumps is that it’s slow and expensive. Doing JIT VLM OCR allows the agent to filter through the data cheaply, but still preserve accuracy for the context that’s needed for the task. The agent harnesses do two-pass document processing by default using off the shelf-tools: pdf2text as the first pass, and using itself (Opus 5) as the second pass. See the below video where Cowork runs over a bunch of PDFs to answer a question about a benchmark graph in the Kimi k3 paper. The main issues here with the “out of the box” doc processing these agents offer are: * Opus 5 is not the best VLM for OCR. It is also way too expensive at scale and lacks grounding * The OSS tools like pypdf, pdf2text, may not be versatile enough as the first pass. * The agent will write a lot of throwaway code to rewrite things an OCR tool would’ve provided out of the box, like chart processing, bounding boxes, confidence scores, leading to increased cost and speed. We have all the tools within LlamaIndex 🦙 to help any agent do two-pass document processing with higher accuracy and lower cost. 1️⃣ We have liteparse for the first pass - a free/OSS parser written in Rust that’s faster/more accurate than other OSS parsers, and supports 50+ document types 2️⃣ We have LlamaParse for the second pass - an agentic document engine that uses VLMs+harnesses to achieve SOTA in accuracy and cost across various doc parsing and extraction tasks. It can be called from any agent harness as an MCP or skill. It takes in page numbers as input, so that the agent can choose to run LlamaParse over a subset of the doc instead of the full doc as a “zoom-in” pass. Come check it out! LiteParse: LlamaParse: All the relevant docs, including MCP, are here:

Jerry Liu

23,239 просмотров • 1 месяц назад

I think we're about to stop editing videos. Not because AI edits faster. Because it's starting to understand what you're trying to say. I think we're about to stop editing videos. Not because AI edits faster. Because it's starting to understand what you're trying to say. I found a tool called Pexo. I found a tool called Pexo. I used it to create a 45-second collage animation of two pugs climbing Mount Fuji. Think of it as a playful concept for a pet travel brand. Instead of opening a timeline and dragging clips around, you give it a URL, document, or source material and tell it what you want. Then you talk to it. It builds the video. Instead of opening a timeline and dragging clips around, you give it a URL, document, or source material and tell it what you want. Then you talk to it. It builds the video. If you want changes, you don't re-edit the timeline. You just ask. That's the interesting part. If you want changes, you don't re-edit the timeline. You just ask. That's the interesting part. • Turn URLs into videos • Use docs or source material as input • Refine videos through conversation • No editing experience required • Built for product launches, explainers, and brand content • Keep iterating until it's ready to publish • Turn URLs into videos • Use docs or source material as input • Refine videos through conversation • No editing experience required • Built for product launches, explainers, and brand content • Keep iterating until it's ready to publish We're moving from editing videos to directing AI. That feels like a much bigger shift than most people realize. We're moving from editing videos to directing AI. That feels like a much bigger shift than most people realize. Get free credit by using this code: B4QFWK Try Pexo here:

Hasan Toor

174,034 просмотров • 27 дней назад