We are excited to announce that Matchstick Ventures Fund III has completed an investment in Freeplay. Freeplay helps product teams ship better products with LLMs, by giving people the power to prototype faster, test with confidence, and optimize features for customers.

‍

With the recent rise in the capabilities of AI, including releases from OpenAI and ChatGPT, there has been an explosion in the number of software companies using Large Language Models, or LLMs, to enhance their products and customer-facing features in ways that felt impossible not long ago. Not only is every established software company looking to integrate AI into their existing products, but there is also a new generation of startups building AI-native applications. 

‍

But while AI has moved from the domain of researchers and experts into product development teams, most engineers and product managers lack experience and need help using AI in production. It’s hard to ship software powered by LLMs with confidence, especially when results are non-deterministic and there’s an infinite set of models and data pipelines, edge cases to test for, and multi-dimensional ways to evaluate the quality of LLM responses. 

‍

How can software teams prototype, test and iterate on these new systems quickly and consistently, so they can deliver better quality experiences to their customers? This creates an opportunity for companies and products to help teams put the processes and tools in place that they need to build with LLMs effectively, without having to become experts themselves. 

‍

Freeplay offers a better way to build with LLMs and provides prompt engineering, testing, and evaluation tools for product teams. Freeplay brings engineers together with product managers, designers & domain experts, who can use Freeplay’s suite of features to collaborate to deliver increasingly better customer experierences with LLMs..

‍

Freeplay provides tools for both software engineers and product managers to iterate on LLM prompts, run batch tests, evaluate outputs, and monitor what’s happening in production. Teams can compare different versions of prompts and different models, log and inspect results, test at scale, and automate it all to compare it to prior examples while also collaborating on future iterations to improve outcomes. 

‍

With Freeplay, product teams can gain confidence their LLMs are working as expected in production, improve customer experiences, and limit downside risk from false or embarrassing responses. 

‍

The Matchstick and Freeplay teams have known each other for more than a decade, and we watched their work at Gnip & Twitter to build and scale their enterprise data business from nothing to nearly $400 million in revenue. The Freeplay team has deep experience around building developer tools and interfaces, and they are positioned well to win in this emerging market.