šŸŒŸ Pezzo is open source. Show your support by giving us a star!

Pezzo v0.5: Dashboards, Caching, Python Client, and More!

This version brings a lot of new features and improvements to Pezzo. We're excited to share them with you!

Ariel Weinberger
Ariel Weinberger
September 2, 2023
Ariel Weinberger

Before we dive into this version, I want to thank everyone for the amazing feedback we've received since we launched Pezzo. We're thrilled to see so many people using Pezzo and building amazing LLM-powered applications. We're committed to making Pezzo the best LLM platform out there, and we're working hard to make it happen.

Is there a feature you'd like to see in Pezzo? Let us know by opening an issue on GitHub!

What's New

šŸ“ˆ Project Dashboard

We've added a new screen to the Pezzo Console. The Project Dashboard gives you a quick overview of your project's performance. It features several useful features:

  • Timeframe Selector: Select from hourly, daily, weekly, monthly, yearly and even custom timeframes for analytics.
  • Vital Metrics: You can now see the number of requests, cost, average request duration, and success rate. You can even see how they change over time.
  • Useful Charts: We've implemened two charts -one for total requests (as well as errors) and average request duration over time. We'll add more charts in the future.

šŸ’” Have an idea for a useful metric or chart? Please let us know on Discord or via a GitHub issue. We're always looking for ways to improve.

šŸŽļø Request Caching

We've implementing a caching mechanism. This feature can help you save up to 90% of your LLM API costs and time!

Some practical use cases:

  • Development: During development developers tend to go through flows very frequently. This usually involves the same set of LLM API calls with the same input date. With Pezzo, your entire organization can share the same cache, and focus on value!
  • Production: If you're building a support chabot, for example, there are many queries that are highly repetitive. For example, "What is your return policy?" or "What are your opening hours?"

You can read more in the Request Caching documentation.

We're planning to add more caching features in the future, such as semantic caching.

šŸ Python Client

We're excited to share that Pezzo now features a Python client! Here are some useful links to help you get started:

We've also made sure to add a Python copy-pastable code snippet in he console, to make it even easier for you to get started.

Thank you!

We hope you enjoy this version. We're working hard on the next one, which will feature a lot of exciting features and improvements.

If you haven't already, please consider giving us a star on GitHub to make sure more people can benefit from Pezzo. ā­ļø

Try Pezzo Cloud šŸŒ©ļø

Our cloud-based LLMOps platform is now available!