🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Are we at a point now with LLMs where bad RAG responses are almost certainly because your query or retrieval was bad? Seems like most LLMs these days will give a correct answer assuming it has the correct context 99% of the time?

Was a ton of fun getting this to work! If you've never used it, Reflex is pretty cool for creating full-stack apps in python 🔥

Workflows from LlamaIndex 🦙 and OpenAI's new o1 model are the perfect match to create multi-agents. Now, also in LlamaIndex for Typescript. Use our app-creator to generate code using the OpenAI o1 model and workflows. 📄 ts.llamaindex.ai/modules/workfl… 🛠️ github.com/run-llama/app-…

Its alive! Well, kinda github.com/logan-markewic… (This is the first thing I've ever made with rust)

This is a huge upgrade in performance, I recommend trying this out!

Wild idea -- LLM providers give users an API endpoint to count the tokens in your input before running Counting function calls, system prompts, etc. is a nightmare

Using LlamaIndex Workflows to implement an async function calling agent with streaming and dynamic context I'll have to add this to docs soon, I keep scrolling back in my slack history to find this link colab.research.google.com/drive/1wVCkvX7…

LlamaIndex 🦙.TS will have the best document for building RAG using LlamaIndex TS. We gonna built lots of rich demos in frontend shows our modules step by step ts.llamaindex.ai/docs/llamainde…

It's been a huge year for LlamaIndex! Tens of millions of pages parsed, millions of framework downloads, and tons of happy users! Check out our year-end review in the numbers, including a month-by-month breakdown of our biggest feature releases: llamaindex.ai/blog/the-year-…

Workflows in LlamaIndex are great, but I really think a higher-level entry point was missing. That's exactly why I built AgentWorkflows, to give users a pre-built starting point that is (hopefully) generally useful. Check it out!

Integrating the responses API from OpenAI today Below is a valid input to the API -- I miss chat messages already 🥲

Harrison Chase Dude - I like the overall idea of the blog and I'm not expecting a 100% unbiased comparison, but the checkboxes on LlamaIndex 🦙 are inaccurate. We've been working on both high-level agent abstractions and low-level workflow abstractions for almost a year now. ✅ We have a

Are defaults in frameworks bad? Seriously considering removing the openai defaults from llama-index (with an appropriate deprecation period ofc) For example, ada-002 is not great, but changing that default to something else is just such a huge breaking change

People always ask me when to use multi agents, single agents, or hardcoded workflows. The good news is, LlamaIndex has solutions for every case. Here is some excellent thoughts and guidance.