Based between London and San Francisco, Klu, a leading large language model (LLM) application platform for building, evaluating and optimising AI applications, has raised a €1.6 million pre-seed funding round led by Firstminute Capital.
Scouts from top venture capital firms including a16z, Sequoia Capital, Craft Ventures, and Atomico also participated alongside angel investors from Superhuman, IBM, Productboard, and other leading technology companies.
The new capital will fuel Klu’s mission to power the rapid emergence of dedicated AI engineers and teams across organisations. Klu offers a suite of integrated products enabling companies to quickly go from idea to optimised LLM-powered features built on large language models like GPT-4.
“We are thrilled to back Klu as they enable the next generation of AI-powered software. With Klu, knowledge workers can spin up AI-based internal workflows to automate repetitive tasks and embed AI workflows into their products for customers to experience. This is a major product challenge which is why Stephen and Stefan with their prior roles at Productboard and Skyscanner are the right duo to solve this problem,” commented Sam Endacott of Firstminute Capital.
Co-founded in 2022 by Stephen M. Walker II (CEO) and Stefan Sabev (CTO) and having launched its closed beta launch in March, Klu has been serving over 500,000 requests weekly, experiencing an impressive average weekly growth rate of 10%. Klu, which has offices in London and San Francisco, predominantly, but not exclusively, counts B2B SaaS companies as clients, with customers including Colab, Zavvy, and Productlane.
“The layoffs of late 2022 shifted modeling work from data scientists to product teams consuming AI APIs. These new AI Teams use Klu to launch LLM-powered features in days that outperform old models built over months,” said Klu CEO and co-founder Stephen M. Walker II.
Founded in 2022, Klu’s vision aligns with the accelerating trend of high-growth startups forming specialised squads focused solely on identifying AI use cases and bringing them to market. However, these nascent teams often struggle with infrastructure complexity, slow iteration cycles, and a lack of dedicated tooling.
Klu overcomes these hurdles through products like Klu Studio for quick prototyping, Klu Engine for accessing leading AI models like GPT-4 and Llama 2, Klu Eval for running AI-powered evals on new iterations, and Klu Optimise for leveraging real-world usage data to continually improve outputs with fine-tuning.
A few early adopters attempted to build similar platforms on their own, but ran into challenges, setbacks, and maintenance issues. Learning from these pain points, Klu optimised its solution so now any business can deploy, observe, and optimise an advanced LLM-enabled feature in under 10 minutes – a process that would take months or quarters when building internal ops systems.
Klu supports all major LLM providers (including OpenAI, Azure Openai, GCP Vertex, AWS Bedrock) while also offering seamless connections to data sources, including 24 file types. Customers can then effortlessly integrate their Klu-created LLM apps with various partners, including Salesforce, Atlassian, Supabase, Asana, Intercom, Zendesk, Slack, Notion, and Google. Builders know that getting to production is only the beginning – Klu customers receive continuous observability insights and evaluation of prompt changes as AI engineers update features and prompts.
