Skip to content

Build journal3 min read

The tech works. Now we need customers.

By Cédric · Published · This entry recounts 4 November 2026.

We have videos we can translate and lip-sync. My cofounder has joined me in Thailand. Today's question is how to turn those technical tests into conversations that could help us find our first paying customers.

My cofounder has arrived

My cofounder has joined me in Thailand. I'll introduce him in a video in the next few days. For this update, I wanted to share the next thing we're working on together: finding our first paying customers.

Over the last few days, we have tested video translation and lip sync. We now have examples to show people. Today's plan is to put those examples in front of creators whose content could be useful in another language.

The episode for this entry
56.5 seconds

Recorded in French; AI-dubbed and lip-synced into English, with localized subtitles and graphics.

Same translated audio in both versions.

The English and Spanish videos disclose the original French recording at the beginning. The player pauses at the same moment when switching between the before and lip-synced editions.

Download subtitles

Start with people who already teach

We're looking for people who already have an online course that works in its current language. The idea is to offer them a way to translate that teaching for another audience.

In the video, I mention Instagram as the main place to begin, along with TikTok. Skool is another place I want to look for course creators, then find the social accounts where they share their content.

Their own video is the demonstration

The approach is simple: find a creator's course and social accounts, choose an existing Reel or a recent post, and prepare a translated version. Then send a direct message with something concrete to discuss.

The question I want to test is whether seeing their own content in another language makes the idea easier to evaluate. A creator can hear the voice, look at the subtitles and lip sync, and decide whether this could be useful for their course.

About 50 contacts. About 10 demos.

For this first test, the plan is to contact around 50 people and prepare translated, lip-synced examples for around 10. Producing a demo has a cost, so I want to keep the experiment small enough to carry out and learn from.

Those numbers describe the plan announced in this update. I haven't reported outreach results, bookings or new customers here. The next update will be a chance to share what actually happens.

This episode is a demonstration too

I recorded today's update in French. You can watch the original or choose English or Spanish, with AI dubbing, localized subtitles and graphics, and lip sync. The translated editions disclose that the recording was made in French.

The player also lets you compare each translated edition with the version before lip sync: the same translated audio, with my original French mouth movements. Watch, listen and compare. That's the kind of concrete example we want to put in front of course creators.

Build journal4 min read

Eight minutes of video. A new problem: cost.

2 November: a real yoga lesson, translated from English into Mandarin. The result gives us something to show—and a cost problem to investigate.

Build journal3 min read

Four languages. One feed. The next problem.

1 November: the translated videos exist. Now I have to work out how people should find the version they understand—and how to keep the publishing manageable.

Build journal3 min read

I stopped using language as an excuse

31 October: a camera, a French recording and a question. Could I speak naturally without giving up on an international audience? We put the idea to the test.