Amberscript, an Amsterdam-based AI speech recognition company and Metrix Partners client, raised €8.65M (about $10M) in a Series A in November 2021. Endeit Capital led the round. Metrix supported Amberscript on the raise.

What Amberscript does
Amberscript builds software that automatically turns audio and video into text and subtitles using speech recognition. Customers can use the fully automatic output or order a version checked by professional transcribers. Users include media companies, universities, researchers and businesses that need accessible content.
Accessibility rules are a growing driver of demand. Public bodies and many companies must provide captions and transcripts so that people with hearing impairments can access their content. Subtitles also increase engagement on social media, where much video is watched without sound.
The Amberscript Series A
Endeit Capital is a German-Dutch growth investor founded by former executives of the media company Endemol. Its backing validated the potential of Amberscript’s technology in the media industry. According to EU-Startups, the funds support product development and the goal of becoming the leading subtitling provider worldwide.
Metrix Partners’ role
Metrix supported Amberscript as a client through the Series A. Preparing a software and services company for a growth round involves:
- Separating software revenue from human-verified transcription services.
- Gross margin by product, including speech recognition compute and transcriber costs.
- Customer cohorts across self-serve users and enterprise accounts.
- A financial model linking product investment to revenue growth.
- Support through investor due diligence.
Five lessons for investors in AI speech technology
- Hybrid models need split margins. Automatic and human-checked services have very different gross margins. Report them separately.
- Accuracy is the product. Word error rates by language decide which markets the company can win.
- Languages expand the market. Each new language opens customers but needs data and model work.
- Self-serve and enterprise behave differently. Individual users churn more; enterprise contracts are larger and stickier.
- Compute costs matter. Processing hours of audio uses significant compute. Cost per audio hour should fall over time.
What AI founders should prepare for a Series A
- Revenue and gross margin by product line, reconciled to the accounts.
- Cohort retention for self-serve and enterprise customers.
- Unit costs per hour of audio, split between compute and human work.
- Accuracy benchmarks by language.
- A hiring plan tied to product and sales milestones.
Why accessibility drives growth
European accessibility legislation requires public sector websites and a growing range of private services to make content accessible. Universities record lectures, broadcasters publish more online video, and companies run webinars and training. All of that content needs captions. Automated transcription brings the cost down far enough for captioning to become standard rather than an exception, which expands the market well beyond traditional subtitling.
Metrix Partners across Europe
Metrix works with clients across Western Europe, including the Netherlands, Germany, France, the UK and Spain. Other clients in AI and data include YZR and techspert.io. Amberscript is one of several growth-stage companies Metrix has supported through a Series A.






