Today, we're excited to announce $5.6M in new funding and the full general availability launch of Freeplay, including our new free tier. Now anyone can access the same powerful tools for prompt management, experimentation, evaluation, observability, and data review that have become mission-critical to leading tech companies ranging from top startups to the Fortune 100.

This new funding round led by Renegade Partners brings our total amount raised to $8.9M. It’s already allowing us to deliver even faster on our vision: helping development teams build great AI products and agents.

Read on to learn more about what we’re building, our funding, and what’s next for Freeplay. 

And be sure to check out the demo video at the bottom to see how you can get up and running testing prompts with your own data and custom evals in less than 5 minutes.

Every Software Team Is Now An AI Team (Or Will Be Soon)

We started building Freeplay in late 2022 because we saw two clear trends taking shape:

1. Every software team is becoming an AI team. If you’re building a software product that doesn’t yet incorporate AI, it likely will soon.

2. Prompts are the new code. The hottest new programming language is English, and that’s opened doors for non-engineering domain experts to contribute to building better software — even if they don’t write code.

At the same time, those trends bring new challenges.

Much of the industry is learning together how to do prompt engineering, how to measure and evaluate the quality of complex model outputs, how to run an ongoing and rigorous experimentation process, and how to monitor non-deterministic production pipelines. The rapid rise of agents only increases the complexity involved with all those things. 

Given the accelerating rate of change in AI development tactics and functionality, every company will need to figure out how to address these challenges, especially as they get serious about running AI products in production.

A Tangible Difference

Freeplay offers product development teams a true end-to-end, enterprise-grade solution to these problems. It's had a clear impact for our customers:

• A Fortune 100 tech company transformed their observability and iteration workflows. From a Principal ML Engineer: “I can’t even express how much easier it is to debug our system now that we have completions and traces flowing into Freeplay. Y’all have built a truly excellent product.”

• A large fintech company completely overhauled their evaluation process, achieving a 10× efficiency gain in processing data. From their AI Ops lead: "Freeplay gives us the autonomy to look at accuracy at a larger scale" than was possible before.

• Help Scout gained the ability to ship AI features faster, and the confidence they needed to shift their entire pricing model away from traditional SaaS. From Luis Morales, their VP of Engineering: "The time we're saving with Freeplay is invaluable… Working with the Freeplay team is one of the best tech partnerships I have participated in."

• Postscript described making a major backend transition between LLM APIs that wouldn’t have been possible otherwise, and they can now handle similar changes much more easily. “The service we’ve gotten is amazing, the platform is flexible to meet a lot of needs, and the downstream result is that we can just ship products with confidence.”

How Freeplay Works

Freeplay provides the full end-to-end platform for AI teams to create the data flywheel they need for continuous product optimization:

• Experiment with prompts and models in an intuitive playground, using your own models

• Define and run custom evaluations that matter for your specific product, for both online logs and offline tests or experiments

• Launch batch tests from your code or the browser to quickly make both quantitative and qualitative comparisons across prompt, model, tool, and agent versions

• Observe production and inspect traces with instant, full-text search to filter across all AI interactions

• Identify and label good and bad outputs to create datasets for testing and fine-tuning

• Track and manage inference costs across models and providers

You can learn more about each of these core features and see demo examples on our blog.

We’re building toward a future where Freeplay automates the hardest parts of the AI product optimization process for you — making it even easier to deliver high quality products, and letting you put more energy into shipping new value to customers. For instance, we already help teams automatically create custom evals and align evals to human experts’ expectations. More on this soon! Watch this space. 👀

On the developer experience side, Freeplay’s SDKs and API are designed to give you full control and minimize dependencies, while still providing powerful capabilities:

• Our SDK doesn't force you to route through a proxy or require you to adopt a specific framework. You can use any LLM API, write your own code in any language, and have the option to integrate with popular frameworks like the LangChain, Google ADK, and OpenAI Agents SDK.

• Freeplay won't break your production app if our service is ever unreachable. Check your prompt & model experiments into code and log results asynchronously. We don’t touch your business logic or LLM request.

• We offer true deployment flexibility. Use our fully-managed cloud, or deploy in your own VPC across any major cloud — on Google Cloud, AWS or Azure.

And for enterprises, we've prioritized what matters so that it’s easy for the largest companies in the world to work with us — including SOC 2 Type II certification and GDPR compliance, the option to sign BAAs for HIPAA, SSO and granular role-based access controls, and more.

Building Freeplay’s Future

Our latest funding round was led by Roseanne Wincek from Renegade Partners, with participation from existing investors Sarah Guo at Conviction and Natty Zola at Matchstick. We're also thrilled to welcome new industry leaders and funds, including Next Frontier Capital, PWV (led by Tom Preston-Werner, GitHub cofounder), Vermillion Cliffs Ventures, and a deeply experienced group of new individual operators and investors with backgrounds at companies like Atlassian, Roblox, Dropbox, Scale AI, GitHub, and more.

This new round has already allowed us to go faster on our roadmap, and to expand our Colorado-based team.

PS: we’re hiring! Reach out if you’re interested, we’d love to talk.

Get Started with Freeplay Today

If you’re curious to see how Freeplay could help you today, check out this demo to see what it looks like to get started using your own custom evals to test prompt and model changes at scale.

Whether you're just beginning your production AI journey or looking to take your production quality and workflows to the next level, Freeplay provides everything your team needs to build faster with confidence.

Ready to see Freeplay in action? Sign up now to start building better AI products, or contact us to learn more about how Freeplay can support your enterprise needs.