Lotti is an open-source app for journaling, audio notes, photos, tasks and time tracking, synced end-to-end encrypted between your devices via Matrix. Or, if you prefer enterprise lingo, a distributed and open-source system of record for your memories. Lotti has an agentic layer where you get to pick where inference runs, including on your own device if powerful enough. Out on Linux via Flathub now (7500 downloads to date) and (new) also on Android (which you can sync if you have a Matrix home server running). macOS is available via GitHub releases, iOS will follow. I’m working on a hosted sync service, but there is no ETA for that yet.
I started working on this app about 10 years ago, inspired by my late grandmother. We went through her photos, she was 92 at the time, and she could not remember the year of a trip to Iceland, and with who. After some frustration and palpable sadness, she found notes about all her trips, and the sheer delight and happiness that came from her memory being refreshed will forever stick with me, and I decided that day that I needed something similar, except not on paper. Initially, Lotti (named after her) was a travel journal app.
Fast forward 10 years and it's now a distributed system of record for my memories, all kinds of memories, where all data is synchronized between my devices, without any server-side storage of anything unencrypted, using a Matrix home server to relay encrypted messages between devices, and where there is an agentic layer so LLMs can help me make sense of my memories, but they cannot alter my memories without explicit approval by me.
On top of the system of record, there are different applications that all have to do with memories in the broader sense. The one I use most is task management, with time recordings and audio notes about my tasks and learnings. So far I’ve completed over 4000 tasks with it, building a knowledge graph about the related categories and projects along the way, providing ample context for new tasks in the same categories and projects. Accompanying agents help me move tasks to done.
You can now even create prompts for agentic coding in here. I create a task in Lotti, speak everything I want to get done (preferably while out on a walk), then compile the prompt for the agentic coding tool using Kimi K3. Lotti then watches resulting PRs and checks off automatically what was done.
On a very powerful computer, you can even run inference locally so no data leaves your device, and I think eventually it will be feasible to do everything locally. The second best option for me is to use a GDPR-compliant European inference provider with a Zero Data Retention policy.
New features include different agents for different tasks, such as long running agents for helping me achieve my goals, track my relationships (with a briefing for next time I call my aunt, for example), and for planning my days. Lots happening in this area, this is an active area of development.
I’m looking forward to moments like with my grandma, bringing back pleasant memories (probably decades from now), and I’m making it my mission to help you achieve the same. By design and principle, there will never be any server-side storage of your precious memories in any readable form. All there is, if you opt in for sync, is ephemeral storage of encrypted messages for information exchange between devices, deleted again after sync if properly configured, but not readable by anyone outside the trusted circle of devices in any case.
Data layer and sync have been hardened over the past 5 years or so, and I’m currently managing close to 300K entries in them, fully and reliably synced between my devices, and with over 11K hours tracked. So far it’s scaling well.
Looking forward to feedback. I’m happy to help you set it up, and learn how to make onboarding easier. Everyone deserves to own their memories.
matthiasnehlsen•39m ago
While I do want this app to become useful for many, many users, I absolutely do not want the responsibility for your data, and the only sensible way to achieve that is by design, that even when I operate a sync service, it would not be possible for me to read anything, and equally important, it would not be possible to subpoena any data either. I’m thinking about the Signal incident where a grand jury subpoena to get all data of a particular user was basically responded to with “yes, the user was created at this date and this is when they were last seen” but zero information beyond that was technically possible to provide. That’s the goal for the sync service that I will provide one day. Currently, I know absolutely nothing about any user, all I see is the aggregate install counts on Flathub, and how many users have downloaded by country.
matthiasnehlsen•38m ago
I wanted to post this on Show HN, but that was restricted for me so I'm posting it as a regular thread.
matthiasnehlsen•40m ago
I started working on this app about 10 years ago, inspired by my late grandmother. We went through her photos, she was 92 at the time, and she could not remember the year of a trip to Iceland, and with who. After some frustration and palpable sadness, she found notes about all her trips, and the sheer delight and happiness that came from her memory being refreshed will forever stick with me, and I decided that day that I needed something similar, except not on paper. Initially, Lotti (named after her) was a travel journal app.
Fast forward 10 years and it's now a distributed system of record for my memories, all kinds of memories, where all data is synchronized between my devices, without any server-side storage of anything unencrypted, using a Matrix home server to relay encrypted messages between devices, and where there is an agentic layer so LLMs can help me make sense of my memories, but they cannot alter my memories without explicit approval by me.
On top of the system of record, there are different applications that all have to do with memories in the broader sense. The one I use most is task management, with time recordings and audio notes about my tasks and learnings. So far I’ve completed over 4000 tasks with it, building a knowledge graph about the related categories and projects along the way, providing ample context for new tasks in the same categories and projects. Accompanying agents help me move tasks to done.
You can now even create prompts for agentic coding in here. I create a task in Lotti, speak everything I want to get done (preferably while out on a walk), then compile the prompt for the agentic coding tool using Kimi K3. Lotti then watches resulting PRs and checks off automatically what was done.
On a very powerful computer, you can even run inference locally so no data leaves your device, and I think eventually it will be feasible to do everything locally. The second best option for me is to use a GDPR-compliant European inference provider with a Zero Data Retention policy.
New features include different agents for different tasks, such as long running agents for helping me achieve my goals, track my relationships (with a briefing for next time I call my aunt, for example), and for planning my days. Lots happening in this area, this is an active area of development.
I’m looking forward to moments like with my grandma, bringing back pleasant memories (probably decades from now), and I’m making it my mission to help you achieve the same. By design and principle, there will never be any server-side storage of your precious memories in any readable form. All there is, if you opt in for sync, is ephemeral storage of encrypted messages for information exchange between devices, deleted again after sync if properly configured, but not readable by anyone outside the trusted circle of devices in any case.
Data layer and sync have been hardened over the past 5 years or so, and I’m currently managing close to 300K entries in them, fully and reliably synced between my devices, and with over 11K hours tracked. So far it’s scaling well.
Looking forward to feedback. I’m happy to help you set it up, and learn how to make onboarding easier. Everyone deserves to own their memories.