Local-first extraction
What comes out of LinkedIn or your saves is pulled locally, with throttling, for your own use. Pile captures what you already saved, nothing more.
Why Pile Read less,
Read less,
remember more.
We read a lot. RSS, LinkedIn saves, X bookmarks, newsletters, GitHub stars, YouTube watch-later, Reddit saved. The content piles up, we never come back to it. Pile captures the signal where it lives, summarizes it with your AI, and files it by theme.
The problem A reader aggregates.
A reader aggregates.
Pile understands.
Stacking feeds solves nothing: you just swap one inbox for another. Pile summarizes each item, groups it by theme, and brings it back when you need it.
The detail that changes everything: on LinkedIn, the most useful link (the study, the repo, the tool) often lives in the comments, not the post. Pile extracts those hidden links. You stop missing them.
- Automatic summary, score and tags on every item
- Hidden links from the comments, extracted
- Search by words or by meaning, plus tidy vaults
“The goal is not to read everything. It is to find, at the right moment, the precise thing you half remember.”
Three rules
we don’t negotiate.
No ads. No reselling. No model trained on your content.
Your AI, your key (BYOK)
Plug in your own key. It is encrypted at rest (per-workspace AES-GCM envelope). Your summaries run on your model, not on a black box.
Export, no lock-in
Markdown and JSON, available anytime. The export is an open exit door, never a feature you have to earn.
Made in the Merry Forge studio.
Pile is built by Merry Forge, a personal AI studio. Solo, with no inherited tech debt: an open-source-friendly alternative to LinkedMash, with no forced SaaS and no nested subscriptions.
Ready to read better,
retain more?
Connect your first source in seconds, summarized and organized for you.
Get started