Agents are the following massive factor in AI. Some outline these “brokers” in another way from others, however the normal thought is, they’re AI-powered instruments that may carry out duties autonomously.
The agent hype has reached a fever pitch, however one startup was comparatively early to the sport: LlamaIndex. Founded by former Uber analysis scientists, Jerry Liu and Simon Suo, in 2023, LlamaIndex permits builders to construct customized brokers over unstructured information.
“LlamaIndex began as a toy open-source venture in November 2022,” Liu instructed TechCrunch. “I turned deeply thinking about understanding how giant language fashions (LLMs) might be used on prime of proprietary information exterior their coaching set, and constructed an preliminary set of instruments enabling builders to index and embrace information of their LLM apps.”
Utilizing LlamaIndex’s open-source software program, which has racked up tens of millions of downloads on GitHub, builders can create customized brokers that may extract data, generate studies and insights, and take particular actions. LlamaIndex offers information connectors and utilities like LlamaParse, which transforms unstructured information right into a structured format that can be utilized for specific AI functions.
Whereas there are different open-source frameworks to construct AI brokers on the market, LlamaIndex is differentiated by its suite of knowledge ingestion, information administration, and information indexing and retrieval options, Liu stated. It will probably join information from information like PDFs and PowerPoint shows, in addition to apps equivalent to Notion and Slack, with an agent.
Salesforce, KPMG, and Carlyle are among the many corporations utilizing LlamaIndex at this time, Liu stated.
“All of those competing options remedy particular issues at completely different components of the generative AI stack, however then it’s the developer’s accountability to piece collectively fragmented options to create a working agent,” Liu added. “This can be a vital ache level that hampers transport brokers to manufacturing. LlamaIndex made it our mission to ship probably the most safe, correct, and easy-to-use platform for constructing end-to-end data brokers.”
LlamaIndex’s subsequent chapter is an enterprise service constructed on prime of the corporate’s open-source choices. Known as LlamaCloud, it lets clients create cloud-hosted brokers that may work with and manipulate unstructured information in a wide range of codecs.
LlamaCloud might be deployed by way of a software-as-a-service set up or in a digital personal cloud, and comes with options together with role-based entry management and single sign-on, Liu stated.
Partly to assist fund LlamaCloud’s growth, LlamaIndex not too long ago raised $19 million in a Sequence A funding spherical that was led by Norwest Enterprise Companions, and noticed participation from Greylock as effectively. The brand new money brings LlamaIndex’s complete funding raised to $27.5 million, and Liu says that it’ll be used for increasing LlamaIndex’s 20-person crew, and product growth.
“We have now ample runway to take us by preliminary industrial growth of our platform,” Liu stated. “We’re betting on a future the place builders play a giant function in delivering GenAI functions inside the enterprise.”