New Features

1. Online Dashboard via memex.garden/dashboard

An online dashboard and a full design upgrade for the search interface. If you’re looking for an alternative to NotebookLM this might be it!

Watch: https://youtu.be/X5_h-lPpavo

!!! IMPORTANT NOTE: FOR THE SEARCH IMPROVEMENTS TO TAKE EFFECT, GO TO SETTINGS > RECREATE INDEX

2. 5-10x Price drop for vector search

We were able to swap out our initial solution for providing semantic search (supermemory) and use our own vector embeddings now with the best model on the market, effectively enabling a 6x price drop and as a result the ability to semantically index all content you save which drops the price from ~12-20 credits per fully indexed page to 2-3credits. (really depends also on how long the content is you save. (we do some summarisation and analysis before the indexing to keep costs as low as possible while aiming to keep quality of retrieval high)

Quality of Life improvements

Copy AI chat output as Markdown, Rich Text and HTML

Added dedicated filters for X bookmarks, like and retweets

Click-through to references

Opens highlights and chat references now automatically in page if the extension is installed (highlights also without)

Mini Player for youtube search and AI chat references

Input needed:

Obsidian Integration

We’ve started on planning out the Obsidian integration and already got some good feedback from some of you on how that should look like.
It would be great to get more input on the workflows you want to see.

This is what we got so far in priority:

• Inline display of dynamic tables representing custom search queries (e.g show me a list of all the stuff I tagged with “nuclear-energy” on my article about the state of Nuclear in 2026.

• Searching your Memex from inside Obsidian to directly reference content

• Exporting saved pages and annotations

I’d appreciate the input in Obsidian Integration

Stability and reliability

Done

• Comprehensive test coverage of everything related to saving content

• Error message for any errors related to saving content so you never end up in a situation where things might not work and in the worst case you wouldn’t even know.

• Improved Auth and pricing flows

General Note

I have to be real with myself that in the last iteration of Memex one key mistake we made was shipping too fast, always in a rush to get the next feature out, thinking it would unlock new user growth. The result: Every new update broke things left and right. The code became a mess.

…basically against every startup advice for a good reason.

In this iteration our key focus is: Limit the amount of details, get every detail perfect.

In practice this means: Shipping quality over shipping speed, automated tests for everything we can.

There is this great video by Jack Dorsey talking about this:

https://x.com/mrexits/status/1908248564901568985

That being said: We are still early in this Memex rebuild and things might have unexpected bugs and some things are not testable. So please let us know if you run into issues.

It’s our #1 priority to squash those pesky bugs so do let us know via the chat and we’ll fix them within minutes or hours, no days.

Whats next:

1. Make search great

In the theme of delivering the best quality and not just move on to the next big feature we can we will now focus on improving the search quality which will have impact on both the retrieval inside the app, but also for the MCP server, the upcoming Obsidian integration (search from within Obsidian), mobile search and API search access.

2. Obsidian Integration

Follow updates and add your feedback here: Obsidian Integration

3. Mobile app

Follow updates and add your feedback here Mobile app to save and search

See you at the next Update

~Oliver