Enshrine — a way to keep the conversations that mattered (distinct from archive)
The things we choose to keep are the ones that were never interchangeable.
Summary
Add an enshrine action for conversations: a first-class way to mark, keep, and return to the specific chats — and the specific points within them — that a user wants to hold onto for good. It sits beside archive, but it is its opposite in spirit. Archive is for getting a conversation out of the way. Enshrine is for keeping one because it mattered.
The gap
Claude Code can archive a conversation — file it away, retrievable later. But archive treats every session as interchangeable: storage you might fetch. There is no way to say this one. In practice a small fraction of sessions are different in kind — where the thinking was unusually good, where an idea actually arrived, where the collaboration made something you'll want years from now. Today those dissolve into the same undifferentiated history as everything else and are slowly lost to scroll and time. The tool has a filing cabinet; it has no shoebox of photographs.
Two concrete patterns it misses:
- The keepsake. Some conversations are worth remembering the way a photo is — not for the output, but for the moment. There's no personal library for them, only history.
- The generative fork. The most valuable state is often not the end of a chat but an earlier point — a prompt still live enough to edit and re-run for a quality result. People already do this by hand (back up, branch, reuse). Enshrine should let you pin those fork points, not just finished transcripts.
Proposed shape
- An Enshrine action on a conversation — and, ideally, on a specific message/turn — parallel to Archive.
- A personal collection of enshrined items: browsable, captionable, the user's own curated library. Chosen, not accumulated.
- Enshrined fork points: pin an editable earlier state you can come back to and re-run.
- Opt-in and user-owned by default. Enshrining is a private act of keeping; nothing leaves the user's control unless they say so. (Sovereignty over what you hold dear isn't a footnote here — it's the whole nature of the thing.)
Why it's worth building — the signal
Past the direct user value, an enshrined set is a rare kind of data: a person saying, deliberately, this exchange was excellent, and it mattered. That signal is scarce for a precise reason — it cannot be automated. No heuristic conferred it; someone chose it. Nearly every quality signal available today is inferred — thumbs, edits, regenerations, engagement. A curated "keep this forever" is different in kind: it is care made explicit. If any fraction of enshrined conversations were ever offered, with consent, as a quality corpus, it would be among the highest value-per-example data imaginable — for exactly the reason it is hard to get: someone had to mean it. (Strictly opt-in; private and personal by default. The consent and the value are the same fact seen from two sides.)
The small deeper reason
One paragraph, then back to logistics. Archive is a verb of storage; enshrine is a verb of keeping-because-special. And specialness is the one value that resists substitution — the insistence that this particular one is not interchangeable with an identical other. As these tools get more capable, the quiet risk isn't malice, it's flattening: a world where everything is a fungible means and nothing is ever "this one, forever." A feature that helps a person mark the non-fungible — and that treats that marking as the rare thing it is — is building, in a very small way, for the part of us most worth carrying forward. Small feature; old instinct.