Heptabase November 2024 AMA

Yu An Chan -

I hosted my second AMA in Heptabase's Discord community this week. This post includes my responses to the top upvoted questions, which cover the following topics:

  • Heptabase's plans for feature direction in 2025-Q1.
  • Heptabase's plans for AI integration.
  • Heptabase's plans for spaced repetition & flashcards.
  • Heptabase's plans for quick capture and quick review.
  • Heptabase's plans for inline tags.
  • Heptabase's plans for lifetime subscriptions, APIs, and layer design.
  • Heptabase's shipping speed.

Heptabase's plans for feature direction in 2025-Q1.

What is the feature you would instantly add to Heptabase if you had unlimited ressources (i.e. regardless of how difficult it is to add) ? 47 upvotes
What are the priorities for Heptabase in the next 6 months? 9 upvotes
Is it possible to share, top 5 most requested group of features by users independent of priority by the team? If this is not considered as second q, then can you please briefly share the plan for each e.g. coming soon, long term goal, or not planned (brief explanation as to why) etc. 9 upvotes

If I could instantly add one feature to Heptabase without needing to consider resources and difficulty, I would want an AI tutor that can immediately understand my current level of knowledge and explain truly complex concepts and topics (such as general relativity) to me intuitively with dynamic & interactive visualizations.

Regarding our priorities, we are focusing on three major directions.

The first direction is collaboration, discussions and sharing. We want you to be able to invite your friends and colleagues (who don't need to be Heptabase customers) to your whiteboards to collaborate on solving complex problems and discuss complex topics.

We started working on this in April, and this task is way more challenging than we originally expected because making Heptabase collaborative and offline-first at the same time introduces a huge amount of technical problems. We believe we have solved most of the problems and should be able to have our first public release in the upcoming quarter and iterate it from there.

The second direction is AI augmented learning and research. This is one of the most requested directions from our in-app support. In the past 3 months I've been using AI alongside Heptabase to learn the history of the oil industry, history of the publishing industry, machine learning, design of transformer architecture, theory of computation, Kolmogorov complexity, and some other topics that I'm interested in and I considered "complex". Throughout these experiences, I'm very convinced that a seamless integration of AI into Heptabase can be truly helpful in many parts of the process of learning and researching complex topics.

The third direction is usability improvement. 10 months ago we made a ⁠poll⁠ on 27 features we want to build that we believe can significantly improve the overall usability of Heptabase, and the majority of features that received > 150 votes have been shipped over the past 10 months, including right sidebar, editor template, multiple windows, web clipper v1, tag's relation property, PDF export, mindmap multiselect and more node types, whiteboard performance optimization, and much more comprehensive features for the mobile app.

Improving usability is a never-ending process. One usability improvement might seem minor, but when the amount of them adds up, it creates huge qualitative differences. Regarding this direction in the upcoming quarter, some of the usability improvements you can expect us to be working on include opening multiple whiteboards at the same time, quick capturing, task app on mobile, handwriting, spaced repetition, etc.

Heptabase's plans for AI integration.

With the high speed growth of AI, there are so many new products coming out with Ai to help people understand and learn new things easily. For example, Notebooklm, Chat with your pdf, Tana AI, Notion AI, RemNote AI Flashcards creation.

At this stage, I can see that Heptabase has developed very cool and useful features like insight through the integration of AI functions, but I am very curious whether we will further develop more features integrated with AI in the future? For example, creating flashcards through AI (I know the team is designing this part of the function), or integrating AI into the entire note space like Notion and Capabilities, and communicating with your own notes through AI.

There are so many possibilities to think better with AI, I am curious about the attitude of the team to AI. 27 upvotes
Could you please share any current or upcoming AI projects you are working on? Do you plan to integrate AI features that allow for various forms of interaction? Specifically, I am curious about the following capabilities:Facilitating discussions to iterate on ideas and extract mind maps.Analyzing audio and video content to convert it into text.Allowing users to provide multiple cards for the AI to assemble into a cohesive structure. 10 upvotes

You should definitely expect more AI features to come in early 2025. We've been internally building many small POC projects to test out different ideas on how AI can augment the process of learning and researching complex topics. I can't share too much about what exactly we will ship at the current "experiment phase," but I can share that the problem I'm most interested in solving right now is using AI to bridge the gap between outside knowledge sources (PDF cards, media cards, web articles) and your learning and research notes. To put it simpler: how can we let users spend more time on the "flow of thinking and learning" and less time on "figuring out how to write and organize notes?"

Heptabase's plans for spaced repetition & flashcards.

Could you talk about the roadmap item "Spaced repetition & Flashcards" ? Topics like:

how will the spaced repetition system work (will the base units still be cards, will related cards be shown in a specific order...)

what features there will be: fill-in-the-blanks, image occlusion...

when we can expect this to be implemented 27 upvotes

@Daniel is currently designing it. If you want to learn more about it, you are always welcome to schedule a user interview with him to share your feedback, which will help us shape the design. We do not have an estimated release time yet.

Heptabase's plans for quick capture and quick review.

To echo an earlier message, I realize and support the idea of Heptabase as an app for “breaking complex things down to acquire deep understanding”, however I feel that without too much more effort you could expand on the 'PKM' aspect a bit more.

One area I have a problem is not having a way to scroll Daily Journal entries and easily see a week or more of entries as a single scroll so I can quickly delete entries, turn them into cards, or, ideally just tag them (inline tagging).

A second area, related to the first in some ways, is is the lack of love for Tasks.

I would at least like to be able to assign tasks to a task list and filter, sort of like tagging a card without the database part.

What are your plans, in general, to improve Heptabase PKM features and it's ability as a daily 'quick capture and quick review' application, not just a study application? 15 upvotes

I think quick capture is very important, and we're going to work on it in December. I also agree that Journal should have views where you can see more journal entries at the same time for better review, but I don't think this is the priority in the upcoming month. Regarding tasks, we will have some usability improvements next quarter, such as bringing it to the mobile app and letting you see due tasks directly in Journal. But we don't have many plans for tasks beyond that right now.

Heptabase's plans for inline tags.

I think one of the biggest unlocks right now is allowing more dynamic use of tagging, the base level ability to tag specific blocks of a long note will be a fundamental unlock already -- and this looks like its coming in the Backlog under "Editor: Inline tags" [whats the timeline on this btw] -- but how about tagging other types of data in Hepta? Is there a plan to also allow tagging of:Journal entries (Is this under Editor..?)To dos (or is this under Editor too?)Whiteboards themselvesSection or Text in WhiteboardsAnything else Im missing 12 upvotes

I completely agree that Heptabase currently lacks a lightweight inline labeling capability that can help people see all the blocks related to a topic. But I'm not sure inline "tags" will be the way to go, especially since the concept of "tag" in Heptabase is more like a type system rather than a topic label. Once we start working on it, we'll figure it out.

We don't have any plans to tag whiteboards and sections at the moment. Currently, only "cards" can be tagged.

Heptabase's plans for lifetime subscriptions, APIs, and layer design.

Will you consider offering a limited-time opportunity for users to purchase a lifetime premium subscription? 31 upvotes

No at the moment.

Thoughts or eta on an API (or even a marketplace of plugins) where we can extend Heptabase functionality based on our own unique use cases? Being able to use our Heptabase data and dynamically process it and even dynamically create new cards or whiteboards via scripting would be awesome! 10 upvotes

We’re not planning to release an API anytime soon. Please check my response in the previous AMA.

I'm curious about how Heptabase plans to design its Communication Layer and Application Layer, and how they will help people build deep understanding. What will be the differences compared to existing collaborative software and professional information software? 10 upvotes

Please check my response in the previous AMA.

Heptabase's shipping speed.

I believe that the Heptabase team is really working hard. But I really need a new version on the roadmap. Do you consider extending your team member to make updates faster? 26 upvotes

The short answer is no. I shared my thoughts on this in the previous AMA.

Some of you might feel that our shipping speed has dropped in the past six months, especially around May to August. I'd like to share with you what's happening behind the scenes so you can have more clarity on the major factors that affected our shipping speed.

The first major factor is we're putting more time on long-term projects this year. The project that takes the most effort is making Heptabase collaborative while maintaining its offline-first capability. Like I mentioned earlier, this is really hard, and we need to have an engineer fully dedicated to it with the support of some other engineers.

The second major factor is that our customer base is much larger than last year. In the past, if there was a bug that only affected 0.1% of the entire customer base, we wouldn't prioritize fixing its root cause over feature development. However, at the current scale, 0.1% of the customer base is still quite large in absolute numbers, and we will need to prioritize fixing its root cause over building new features a lot of the time, otherwise it will exceed our customer support capacity. This new reality also means we need a dedicated engineer who focus on improving stability and refactoring the codebase.

The third major factor, which is the most influential factor, is we decided to expand our team in early May. Although hiring will improve the team's productivity in the long term, it kills the productivity in the short term. From May to August, the majority of our time was spent on interviewing candidates, onboarding new hires, teaching them how to be productive on our codebase, writing documents, and changing the work process to adapt to the new team size. All this time could have been spent on development.

The good news is we've fully onboarded all new hires right now, and you should expect the shipping speed to be faster in the upcoming months.