Subscribe
Sign in
Active Recall — Summarize anything, forget nothing
You'll likely forget 90% of the content you consume within a week. Recall summarizes online content, connects it in a knowledge graph and resurfaces it on a schedule tailored to your learning curve, helping you remember the information you care about.
250
Replies
@miko_markovic11 our big differentiator is that we build a knowledge graph from any of the online content you consume and save to Recall. Whether its a youtube video, blog post or podcasts - Recall can summarize it and find connections with other content that you have saved before. We are currently the only tool that can do this. The other graph based note-taking apps (obsidian, roam, etc) focus on building a knowledge graph from the notes that you take and is a manual effort. While Recall is focussed on the content you consume and creates the links automatically.
Here is an example of how you can find connections between new content and past content. I have pinned the timestamp from the video above: https://youtu.be/7jOCYJEdp0g?t=80
@paul_richards1 looks pretty cool, where do you see the product going in the future? I think the tagging can be a very useful tool. More thoughts
@paul_richards1 congrats on the launch! in your knowledge graph, highlighting key info and downplaying the trivial stuff would be super helpful. do you have plans to add features for emphasizing important points while minimizing less relevant details?
@paul_richards1 wow never heard of it before but looks very interesting, I love consuming content, a lot of content, can't wait to try it Paul! Good luck on the launch!
I was consciously looking for something like this. I use obsidion and a bunch of extensions to create my own workflow, it works but its not as seamless as Recall. I have been searching for tool like this and there isn't any that works so well. Well done Paul. Also the onboarding experience was very nice. You guys really took care of the details.
@abhishek_saha3 I am so glad to hear this! Would love to hear more. Please send me any feedback you have after using Recall for a while.
@mikestaub quivr is cool but still quite different to usecase to Recall. Recall is focussed on learning/getting the most out of the online content you consume - while quvir is more for chatbot with your documents.
@sentry_co I guess you meant "Total Recall"? This was a top contender. I also thought about making the base version Recall and the premium version Total Recall
@paul_richards1 Yes!!! Themed as Hollywood 80s Schwarzenegger action movie style. Would totally win product of the day / week / month. Probably hit the front page of the verge 😂
Awesome product. Brilliant explainer video! Looking forward to trying out the product over the next few days. Congrats on the launch 🚀
@ohansemmanuel thank you very much. I appreciate the comment about the video. I am not a professional video editor so it really took a lot of effort to get it to that state. And @sankari_nair did a great job.
@paul_richards1 Looks super useful!! For me, the critical missing piece is readwise integration.
@cihat_imamoglu1 thank you! We already have a feature request for it so it will be scheduled for sometime later this year.
idea: good
api: no (as usual, so no chance to switch from another product)
tagging: bad, as most of these kinds of service
about tagging: I don´t get it. the whole product is about using a LLM to understand the context and create summaries, but "tagging" is terrible.
here's an example: https://www.city23.at - what tags do I get? only one: "city23" - wow.
so the LLM is unable to understand that this a digital online radio station ? NO! if you ask chat-gpt for meaningful tags, you will get:
**Keywords:**
1. Digital Radio
2. Music Stream
3. DAB+ Broadcasting
4. Mobile App
5. Vienna Radio
so why is "tagging" always that bad with these products ? you can argue: no need for tagging because the summary has enough content to search for. but in that case I have to say: then remove tagging! either you tag stuff meaningful or you don´t tag anything. bad tagging is the same as not using it all.
usability:
- CTRL+K is not working when a card is opened
- ESC does not close the card
- I can change the card without going into edit mode, thinks can get changed by accident without the user notices anything, opening a saved item always should be open in view mode, that´s a core principal
- search: no AND/OR search
- search: no search by tags (#tag)
- search: no search by timeframe
in general: why should I search for something that is NOT in my cards ? i search here because I want to know if it's in my storage, but instead I get tons of stuff from the internet, WHY ? ;) that makes no sense. i create my own knowledge base to search in there. if I search for something that is not in my cards I want to know that and don´t want to see "something" from the internet and filter with the help of a little dot that indicates, if the item is in my storage or not. if I want to search the net, I will use my preferred search engine, they can do this better as you, so please don´t play google here, makes no sense and so the whole search is very distracting and irritating.
a product like this is only as good as the search. "saving" is only 50%. if you cannot find something here with a very detailed and super fine search, it´s useless.
@michael_bladowski Thank you for your detailed feedback, Michael! We appreciate your insights and are continuously working to improve our Recall.
Regarding the API, we haven't had many requests from users for it. We do offer an export to Markdown feature. Import functionality is on our backlog, and for now, users can paste Markdown directly into the editor.
For tagging, we don't use an LLM. Instead, we rely on schema.org datatypes. For example, we create the tag "city23," which is nested under the tag "website," so anything saved from https://www.city23.at/ will be tagged with "city23" for easier rediscovery. We found that this was the most logical way to categorize information while also giving the user the option to define their own tags.
The primary goal of our product is to resurface content when it is most relevant, based on connections, rather than forcing users to search for it actively.
You can also treat the entities that are extracted as a mean for categorizing your content. For instance, if a persons name is extracted as an entity, you can pin the "person" into the side menu - that way you can always easily access all your cards that mention that person. The idea here is that your content is organically categorized based on what it mentions.
Thanks again for your feedback. If you have any other suggestions or questions, please let us know. 😊
@igor_gligorevic website tags are SEO based, but tags in a knowledge base are human-based. even a child knows that difference. a tool that uses AI but don't uses it for the most simple things - man, that's so damn absurd - LOL ;-))) anyway... good luck with that.....
Is this tool capable of generating summaries for traditional Chinese content as well?
@chris_wang5 yes it is. We currently support the following languages with more coming sometime:
English
German
French
Spanish
Italian
Chinese
Japanese
Korean
Portuguese
Russian
@himanil_gole Thanks! We're interested to hear your feedback after you check it out.
I have been using Recall for more than a year and the improvements have been excellent. Looking back at my notes from March of 2023 when they launched the beta of the browser extension, I can see that I've not used this tool nearly as much as I should have been. I've collected thousands of things in Notion which is fine, but I really could have been collecting those same items in Recall and have a really powerful setup by now. I'm looking forward to reengaging with this tool in a more effective way and leveraging it for all of the workflows that I use.
@just_s Thank you! Do join our discord channel to keep us posted on how your experience goes: https://discord.gg/JF6MEAJe
@esus Thanks Esus! We have an active Discord community. You're welcome to join. We did have a Lifetime deal before, but it's no longer available.
@esus thank you! See the discord channel here: https://discord.com/invite/J9agg...
@marc_b4 thanks for the feedback Marc. Would love to hear in detail on how we could improve the UI?
Thank you for your answer @paul_richards1. In my opinion the color contrast and the font seems a bit off. The app looks "old" reminds me of the 90s web era. And the sidebar with the chips looks a bit overwhelming. Compared to Obsidian for example I expect a more clean and less disturbing interface. Note: that's just my perception.
congrats on the launch. the concept of connecting consumed content in a knowledge graph is quite intriguing. can you share more about how Recall differentiates the resurfacing of content based on the user's learning curve? how does it adapt to different user behaviors or preferences in learning?
@mashy Hi Albert! Thank you for the comment. You can learn a bit more about how the product works here: https://docs.getrecall.ai/active...
@chrismessina spoken like a true VC ;)
We have been using the name Recall long before Microsoft. We have been advised that we could make a claim against them as we are in the same space.
Recall