The Brief
What changed, what keeps recurring, what’s worth a second look. Your home page, written as a short daily brief from your own collection.
Private alpha · now live Save everything. Find it again. Actually use it.
A private home for the articles, repos, videos, recipes, and half-remembered threads you keep meaning to get back to. Your home page reads like a daily brief. Spaces gather around your projects. And Ask answers questions straight from your own material, with citations back to the source. So can Claude, or any AI you connect.
Use with Claude Daily Briefings Save from web & iOS Connect your existing tools
The app model
One collection. Six ways to use it.
Not another folder tree. Somewhere to save things, find them again, and ask real questions of them.
Spaces
Your projects and interests, grouped automatically. A pile of tabs becomes something you can actually work with.
Notes
Markdown, wikilinks, and saved chats sit next to your library of saved content.
Connections
One click from the browser or your phone, plus sync and import from Obsidian, Notion, Karakeep, RSS, GitHub, and a dozen more.
Ask
A built-in assistant that answers from what you’ve saved, with citations back to the source. It drafts notes and connects threads, and it won’t pretend to know things it doesn’t.
Search
Keyword and meaning-based search work together, so you don’t have to remember the exact title to find what you saved.
The reader
It knows what each thing is.
A saved link isn’t just a title and a favicon. Open anything and Knowmarks renders it the way it’s meant to be read: a recipe shows its ingredients, a repo shows its stars, a video plays inline next to its transcript. Seven kinds, one reading surface.
-
Articles & essays
Clean reading with the source hero, byline, and your highlights in place.
-
Video
Play it inline, right beside a timestamped transcript and chapters, without leaving the page.
-
Recipes
Ingredients, steps, times, and nutrition pulled out of the page.
-
Products
The photo, price, availability, and specs, saved for the second look.
-
Repositories
Stars, language, license, last commit, and the README right there.
-
PDFs & papers
The document with an outline and page anchors for every citation.
-
Discussions
The original post and the comments that were actually worth keeping.
Agent-ready
Ask from any AI. Always with receipts.
Knowmarks isn’t only a place to save things. It’s a source your assistant can search, cite, and turn into working notes, over a tool surface built for the way agents actually use a knowledge base. Connect it once, then ask from Claude, Cursor, or anything that speaks MCP.
Connect your AI
Remote MCP endpoint
https://alpha.knowmarks.app/mcp Paste it into Claude, Cursor, or any MCP client. Your saves become searchable and citable from inside the tools you already work in, and they stay yours to take with you.
Under the hood
Two bets most tools don’t make.
Agent-native, not agent-compatible
Most tools build for humans, then bolt on an API. Knowmarks treats the MCP server and REST API as peer surfaces from day one. Agents get full CRUD over the same data model the dashboard uses, not some stripped-down afterthought. In the agentic era, that’s the whole difference between an AI that can use your tool and an AI that lives in it.
Spend your time on work, not filing
Every other knowledge tool slowly turns into a second job: tagging, foldering, pruning dead links. Knowmarks does that part itself. It sorts, groups, dedupes, and notices what’s gone stale in the background, and only asks for your call when something genuinely needs it. Your time goes to the work.
Save from anywhere
Bring in what you already save.
Save in a click from the browser or your phone, then pull in everything you’ve already collected elsewhere. Sync your notes, import your bookmark managers, follow the feeds you care about. It all lands in one collection you can take with you.
Save in a click
- Browser extension
- iPhone share sheet
Sync your knowledge base
- Obsidian
- Notion
Bring your bookmarks
- Karakeep
- Readwise Reader
- Raindrop
- Linkwarden
- Linkding
- Shiori
- Wallabag
- Browser bookmarks
Follow sources
- RSS feeds
- GitHub stars
- Hacker News
Ownership
Your data is your data.
A knowledge layer only works if you trust it with everything. So the rules are simple: what you save is yours, it never becomes someone’s training data, and you can walk out with all of it whenever you want.
Never the training set
Your library is never training data. Knowmarks doesn’t train on a single thing you save.
Not tracked, not sold
No ad tech, no data brokers, nothing sold. The only telemetry is anonymous, and it never includes your content.
Take it with you
Export everything in one click. Markdown plus JSON, ready to round-trip into a fresh instance. Leaving is meant to be easy.
Runs on your machine
Local-first by design. Your files, your model keys, your control. You can run the whole thing on your own machine; the cloud is just convenience.
Inside Knowmarks
Built around the way saved things actually age.
Some links are quick references. Some become projects. Some become notes. Some were only useful for ten minutes. Knowmarks knows the difference.
The projects forming inside your saves.
Spaces show the topics and projects taking shape, with counts, descriptions, and enough structure to act on. Some you name yourself; others the system suggests.
A pile of tabs, turned into something you can work with.
Open a Space and it reads like a collection rather than a folder. Every item shows its source, its kind, and the actions that fit it.
Everything you’ve saved, filterable by kind.
Your whole collection in one place. Filter by kind (articles, repos, videos, PDFs, recipes, products) and open any of them in the reader.
Alpha test stage
Private by design, rough where it should be.
The hosted alpha runs on a private server with separate user data stores. It is meant for a small group of testers, not a public SaaS launch.
The main question is whether the daily act of saving becomes more useful once the Brief, Spaces, search, and Ask have enough of your real material to work with.
Join the alpha
Good testers save too much and remember just enough to be annoyed.
We are looking for people who already save links, docs, repos, videos, or notes somewhere, and who can tell us where Knowmarks fits or falls over.
Email [email protected]No newsletter. No marketing drip. If you’re invited, we’ll ask a few questions about where you save things now and what tools you can’t live without.