What we started with
The first job is getting the source into a form you can read. Articles need the page clutter removed. Podcasts and videos need readable transcripts. PDFs need their text put into reading order, which can be difficult when the layout is complex.
Then comes the part that made the difference for us: being able to open an explanation, ask a follow-up question and write a note while the source is still in front of you. When you want to listen, the text can be prepared for natural spoken narration.
What we are building around it
One source is enough to start. When you have several related pieces, a notebook lets you gather them with your notes and questions. You can return to a passage without having to reconstruct why you saved it.
We want the help to be useful and easy to check. Explanations and answers can be wrong. You should be able to follow a citation, inspect the source and decide what you think.
The case for comprehension explains the research behind these design choices. It is a rationale for what we are trying, rather than evidence that using Omphalis improves learning.
Help us make the reading better
Try a source you know well. If the extraction misses something, an explanation gets a distinction wrong, or you cannot do what you expected, tell us what happened. A specific example gives us something we can fix.
Omphalis is built by Voxiven. These startup programs support the company.
- Cloudflare
- PostHog
- Datadog
- GitHub
- Auth0
- NVIDIA Inception
Logos acknowledge startup program support and infrastructure used by Voxiven LLC. No endorsement is implied.

