LlamaIndex

In the age of agentic AI, context is everything. But there are so many different forms of context. While we started as a broad framework connecting all sorts of data and context to the model layer, today our mission is hyperfocused on unlocking a very specific but universal form of context: documents 📃📄📑 Today, we have best-in-class technology for parsing PDFs, Office docs, and others to unlock and extract context for your AI agents. That's it. Next time you're in SF and you wonder, "didn't LlamaIndex use to be a RAG framework? What happened?" this sign on 2nd Street might help 😉 Come bring your hardest, nastiest PDFs, we will parse them with LlamaParse. Sincerely, We Parse Docs LlamaIndex

How do you actually know if your document parser is good enough? 🤔 Most teams can't answer that. They're either: 🛒 Trying to choose a parser and have no real way to compare them apples-to-apples 🔧 Building their own and no real way to tell if it's production-ready Existing benchmarks like OlmOCR weren't built for how AI agents actually consume documents. That's the gap ParseBench fills. But we didn't build ParseBench just for ourselves. So we're hosting a webinar to uncover the methodology that went behind it, to enable all AI teams to evaluate their own document ingestion pipelines with confidence. 🎙️ Inside ParseBench: How to Evaluate Document Parsing for AI Agents 🗓️ May 27 | 9 AM PST What we'll cover: 📊 Where existing benchmarks fall short 🎯 The 5 dimensions that predict parser performance on real enterprise docs 🛠️ How to structure an eval around your own documents and use case If you're picking a parser, building one, or shipping document workflows in production, this is how you start measuring what "good enough" actually means 🔗 Register: https://lnkd.in/eQdxG6pr

You submit a job. It runs. But how long did it actually sit in the queue? How long did processing take? Now you can see both. New: Latency Metrics is now live in LlamaParse. For every Parse, Extract, and Classify job, you now get:   • Queue time   • Processing time   • Total latency Broken down by tier, with a job volume histogram in a new Metrics tab so you can spot patterns over time. Head to your Parse History to check it out. 🦙 Get started with LlamaParse: cloud.llamaindex.ai

We're excited to be an official shoutout at the Google I/O Developer Keynote. We are building the document infrastructure for AI agents, and we plan to integrate even more heavily with both the model layer (Gemini API) and agent harness layer (Antigravity agents) to support all developers within the Google ecosystem.

Financial analysts spend up to 70% of their time pulling numbers out of PDFs. Transcribing 10-Ks into spreadsheets, mapping GL accounts, reconciling trial balances. Every number needs a source. We built a demo that flips that workflow. It's an AI agent that ingests SEC filings, answers questions, and highlights the exact source text on the original PDF page. The whole thing is about 600 lines of Next.js code. No vector database, no embedding pipeline. Just LiteParse for text and bounding boxes, keyword search, and a well-instructed LLM. The full walkthrough covers how the citation system works, why we skipped vector search, the SEC EDGAR integration, and what we'd swap out for production. Read it here: https://lnkd.in/gpjKjfME

We're live at Google I/O 🔥 Lots of exciting features coming to the Gemini API and we're exciting to provide the document infrastructure for Google ecosystem builders.

🚀 The team at Google just released the Agents API, a service for building and running custom agents inside a sandboxed Linux environment, and we built a template that gives these agents access to LlamaParse / LiteParse, enabling them to process unstructured documents automatically 📄⚡ Here’s how it works: 🔹 Configure a Git repository where data and outputs will be stored 🔹 Clone the repository into the agent sandbox 🔹 Install the LiteParse CLI, the LlamaParse SDK, and agent skills to use both 🔹 Prompt the agent with a task and watch it process documents autonomously 🤖 The result? An agent that can work directly with messy, real-world documents using LlamaParse and LiteParse within Google’s new agent runtime. Check out the GitHub repository: https://lnkd.in/gHUrZhsB Get started with LlamaParse: https://lnkd.in/eJ6zujZ5

That's a wrap at AI Engineer Singapore 🇸🇬 Thanks for all the devs that tuned into our workshop and keynote. See you in a few weeks at SF 🌉

A new set of open-weight models is topping the leaderboard for document understanding 🔥 INF AI just released two models: Infinity-Parser2-Pro (35B) and Infinity-Parser2-Flash (2B) that top our Hugging Face leaderboard for ParseBench. Two key insights: ✅ An expanded synthetic data engine over 5 million diverse parsing samples ✅ A novel Joint RL algorithm that co-optimizes multiple complex tasks: document parsing, element parsing, chart parsing, and more. ParseBench is an open benchmark designed to test semantic document understanding over real-world enterprise documents; it has comprehensive metrics over tables, charts, semantic formatting, and more. Come check out the results on ParseBench! HuggingFace 🤗: https://lnkd.in/gaZGbH_a Site: https://www.parsebench.ai/ Infinity-Parser Flash model: https://lnkd.in/gr6qkBBD