The latest updates to LlamaCloud

To build a production-quality LLM agent over your data, you need a production-quality data processing layer. LlamaCloud is that data processing and management layer for your AI knowledge assistants. Since launching a LlamaCloud waitlist last week, we’ve gotten hundreds of signups and published case studies showing how it cuts production development hours by 50%.

On top of that, our team has shipped a slew of new features at a breakneck pace in the past week. We’re excited to highlight these new features that collectively help you set up a chat interface in minutes, increase developer collaboration within your team, and access more data and metadata.

Set up a Chat Interface in Minutes

We are releasing LlamaCloud Chat, which gives you an easy-to-use chat interface over your data. This chat interface is a conversational RAG pipeline built over the advanced retrieval interface that a given pipeline provides, and has out-of-the-box support for streaming and citations - it’s powered by the same DNA as create-llama, our fully open-source set up tool for LLM applications.

The LlamaCloud UI already lets you set up a data pipeline over any data in minutes, and now you get a full-blown ChatGPT over your data in minutes. Besides the chat UI, you also have additional flexibility:

• You can customize metadata filters in the retrieval parameters
• You can view retrieved nodes and their source files
• Besides chunk-level retrieval, you can now do file-level retrieval (more on this soon!)

LlamaCloud is fundamentally a developer tool: with these updates, we enable developers to spend less time on data pipeline setup and iteration, and more time on writing the orchestration logic on top of this interface.

Increased Developer Collaboration

We’ve added organizational features into LlamaCloud, enabling any individual user to create an organization and add other users to the organization. Any user within an organization will have a view of all the organization’s projects and indexes within each project.

This allows your team to have a single-source of truth for your data pipelines. In the past each developer would spend time re-indexing/experimenting with the same sources of data. This feature enables transparency, re-use, and generally more rapid development velocity.

Improved Data and Metadata Access

We’ve made several updates here - we’ve added more data connectors and added features to let you more easily access and customize metadata.

• We added a Notion, Slack, and Jira Connector
• Our Sharepoint connector now natively integrates with user IDs that you can filter for, enabling you to build LLM applications with access control.
• You can now attach metadata to any uploaded file as a CSV - do this through the UI or our API!

Want to see what LlamaCloud can do for you?

Come sign up on our waitlist for access. If you’re interested in chatting about enterprise plans, get in touch.

If you’ve gotten access to LlamaCloud, check out our rich repository of demonstrations and examples on how to build different LLM application use cases.