SIAgentNews.com
← All news
DevelopingBusiness

Generative media startup each::labs raises $7M seed round

The San Francisco startup secured funding led by the EBRD to scale its API platform that provides production-grade video and image generation.

· First reported by The Next Web

$2MMonthly revenue in October

The short version

  • The $7M seed round was co-led by Revo Capital and the European Bank for Reconstruction and Development (EBRD).
  • The platform provides an API for over 600 external models plus proprietary models for image, video, and audio.
  • The company reached a monthly revenue of roughly $2M in October, with over 750 paid customers.
  • A recent system rebuild improved request success rates from 56% to 100% and increased core request speeds by 25 times.
  • The startup launched Lockstep, a model for character swapping in video, priced at $0.09 per second of 720p video.

Why it matters

For businesses, each::labs acts as a unified gateway to the fragmented world of AI models, making it easier to integrate diverse tools into production workflows. Their focus on reliability and high-speed processing addresses the primary hurdles for companies trying to move from AI experimentation to actual commercial use.

Securing growth capital

The San Francisco-based startup each::labs recently announced it raised $7M in a seed funding round. This investment was co-led by Revo Capital, which is Türkiye’s largest early-stage fund, and the European Bank for Reconstruction and Development (EBRD). The EBRD serves as the anchor for Revo’s $100mn third fund.

The company was founded in June 2024 by engineers from Türkiye, including Eftal Yurtseven and a former Insider engineer. This new capital is intended to fuel the expansion of the company's team in San Francisco and support the growth of its physical workspace. The funding will also be used to develop more in-house models and increase the company's sales and marketing reach.

This round follows an earlier outside investment in February 2025 led by Right Side Capital. The startup has managed to achieve significant commercial traction despite having less than $2M in earlier funding. It currently employs a team of about 20 people to manage a platform that serves a wide variety of industries.

Scaling revenue and reach

The startup has seen rapid financial growth over the last few months. Monthly revenue grew about fivefold between August and October, reaching roughly $2M. This growth puts the company's annualized run rate at over $20mn, a significant milestone for a company that is still in its early stages.

The platform serves a diverse range of clients, from production studios and online retailers to ad platforms and consumer AI applications. As of October, the company reported having more than 750 paid customers using its services. One specific example is the ad firm Wask, which uses the platform to generate over 1,000 ad creatives every minute during peak times.

The company's success is tied to its ability to provide a unified API for over 600 outside models for images, video, and audio. By putting these various models behind a single interface, each::labs allows customers to access multiple technologies without having to manage separate integrations. They also train their own proprietary models to supplement this library for specific customer needs.

Technical reliability overhaul

A major rebuild of the platform took place in July to improve the reliability of the system. When the company replayed old customer traffic through the new architecture, the success rate jumped from 56% to 100%. This technical improvement was essential for supporting the surge in usage that occurred in September as consumer apps began moving into video generation.

In addition to the success rate, the rebuild made core requests up to 25 times faster. The system is now designed to handle around a million daily runs. If a generation request fails on one model, the platform is now capable of automatically falling back to another model to complete the task.

These improvements allowed the company to handle the heavy workloads associated with complex generative tasks. One such task is character swapping, which requires maintaining the motion of a clip while changing the person appearing in it. To solve this, the team built a new model called Lockstep, which launched on October 28.

Pricing and studio expansion

The new Lockstep model is priced at $0.09 per second of 720p video. According to each::labs, this price point is roughly one-quarter of what a rival service from Higgsfield costs for the same resolution. By offering a more affordable price for motion transfer, the company aims to make high-quality video manipulation more accessible to a broader range of creators.

Beyond the API, the company also opened each::studio in San Francisco this past October. This physical space is designed for film, VFX, and marketing teams who want to use generative media tools without needing to write any code. It serves as a hub for professional teams to experiment with the technology in a collaborative environment.

CEO Eftal Yurtseven explained that the company's goal is to bridge the gap between having an idea and being able to show it to others. He noted that while people can now show their ideas without drawing skills, they still have to fight with the tools to do so. The company aims to build the layer above the models to make that process seamless for professional users.

What to watch

  • How many more in-house models each::labs will release with the new $7M funding.
  • Whether the each::studio physical space becomes a primary hub for B2B marketing partnerships.

Try it

Search for 'Lockstep' or 'each::labs' to see how they compare their video motion transfer pricing against other market rivals.

Sources and links

First reported by The Next Web. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news