The video is only the beginning.
Watch podcasts with an AI research assistant by your side. As the conversation moves, SeeWith shows what is being talked about: the people, the places, the ideas and the sources.
Context cards are AI-generated and can be wrong.
Contextual discovery
Topics are detected while you watch, and each one brings up to six cards beside the video.
Trusted sources
Every link and image is a provider's own result, with its provider shown. No link is ever written by the model.
Visual understanding
Photographs, diagrams and book covers for what is being discussed, not a wall of text.
A rabbit hole, kept for you
Follow a thread to earlier and later moments of the same episode, and save what matters.
Photograph: Bruce McCandless II, first untethered spacewalk, 1984 — NASA, public domain.
You watch. SeeWith reads along.
Nothing runs until you press Start SeeWith. From then on the companion follows the playhead and reads only the part of the episode you are in, never the whole thing in advance.
- 01
Paste a YouTube link
The episode opens with the player on the left and an empty companion on the right.
- 02
Topics are confirmed as they are discussed
A topic changes only when the conversation really moves on. Each one becomes a stint on the timeline.
- 03
Discoveries arrive beside the video
A visual, the entities named, the sources. Each with its provider, each at its moment.
- 04
Go deeper, keep what matters
The Rabbit Hole links moments across the episode. Saved topics are there when you come back.
Made for the long conversations.
Three-hour interviews, history deep dives, science explained slowly. SeeWith is a companion to the video, not a transcript, a summary or a chat.
Sign in with Google to open an episode.
Sign in to SeeWith