Medal, a startup best known for its video game clipping products, announced that it has raised $13 million at a valuation of $333 million from multiple investors, including Horizons Ventures, OMERS Ventures, peak6 and Arcadia Investment Partners.
The company also announced Highlight, a new cross-platform desktop app that acts as a contextual AI assistant for users, capturing on-screen content and allowing them to ask questions to large-scale language models (LLMs) based on that context.
Henry Gladwin, a partner at OMERS Ventures, said in a conference call with TechCrunch that the venture sees an opportunity to take Medal's core technology and apply it to LLMs.
“Medal's core technology is based on the idea of ​​understanding what's going on on someone's device: video, audio, and what's going on around them. This was originally used for clipping. Now the company is taking that technology and applying it to LLM for prompts, which is a clever use,” Gladwin said.
Gladwin added that he saw Medal as more than just a games company, but a product for capturing the best moments of virtual life, and the Highlights app is a natural extension of that paradigm.
How do highlights work?
For years, companies have been working on building assistants that are helpful to users who leverage on-screen information. Google has been doing that for years with Google Now, Google Assistant, and now Gemini. Apple made inroads in this field last month at the Worldwide Developers Conference (WWDC) with the announcement of Apple Intelligence and the ability to understand contextual information on the screen. Microsoft is also leveraging generative AI with Windows Recall, a feature that helps users find content they've viewed in the past. After the initial announcement, Microsoft decided to delay the release of Recall.
Highlight is trying to bring this to the desktop. In its current version, the app exists as a floating button on the desktop. When you move your mouse over the icon, it captures the content on your screen and passes it as context to various models. You can also ask other questions using various tools such as ChatGPT, Claude from Anthropic, and Perplexity.
Image credit: Medal
Based on the different models, the app will pre-populate some questions to help you get started. The capture is done locally and the app doesn't store any of the content. The company is building its own ChatGPT-like assistant, which may eventually run locally on the device, though it may be less capable than cloud-based models for some tasks.
Beyond on-screen content, documents and system audio memory can also be passed to Highlight as context.To leverage audio use cases, the company is building a local transcription app for meetings, similar to tools like Granola, Limitless, and Krisp.
Building highlights
In a phone interview with TechCrunch, Pim de Witte, one of Medal's co-founders, said the company began thinking last year about how it could leverage Medal.tv's clip-capture technology and blend it with AI.
“We know that recoding activity will be important for operating systems, and we have seen it being done by several major tech companies. We want to provide an open platform that connects users with assistants, models and interfaces,” de Witte said.
He also said that he wants to create an AI app that can be used by people of all ages, even if they are not tech savvy, so the app will give contextual prompt suggestions based on the captcha.
The company is also building an open platform for developers to deploy their own apps on the Highlight platform, similar to the Raycast launcher app for Mac that allows developers to install extensions they create, with a key difference being that Highlight is available for both Mac and Windows.
Image credit: Medal
OMERS Ventures' Gladwin said Highlights has an advantage because it's independent and not interested in pulling into an ecosystem.
The way forward
Medal.tv is a successful product, and the company wants to continue supporting it and adding new features, but it plans to allocate some staff to work on Highlight.
Highlights is free for now, but de Witte hopes to monetize it through an app store model. The startup is also considering premium subscriptions that would give users access to some of its apps and features, such as its local-first model.
The company is also handing out grants of up to $30,000 to developers who will also get access to the Highlight team to develop their app ecosystem.