
Remember This
Where Photos Become Conversations
After more than enough tinkering and exploration of the tech behind Remember This (privacy-preserving photo library analysis), including a few early beta releases, I am looking for advice on taking the right steps toward shipping something that actually brings value to others. If you have a moment, please invest it in reading this roadmap and provide some thoughts: https://rememberthis.ai/blog/updates/2024-12-25/roadmap
Big thanks for any feedback / advice!
With the recent releases of Remember This, the core functionalities are in place:
- Mac Desktop App: Plug-and-play photo metadata analysis environment running on your Mac.
- Local AI Photo Descriptions: Generate descriptions using Local AI — completely offline and private.
- Photo Metadata Exploration: See the when, who, where, and what of your photos in a timeline view.
I wasn't really sure what to focus on next, so I instead focused on creating a proper website that provides the foundation for broader sharing and feedback of the project. Check it out, it sports information about Remember This and even has a rudimentary blog :) https://rememberthis.ai
1 Like
Comment
Included in this version:
✨ Timeline: A new default start screen. See when, who, where, and what you were up to in the past days, weeks, months, and years.
📸 Photos View: View the when, who, where, and what of your photos in the timeline view.
1 Like
Comment
Included in this version:
🔎 Chat to Find Photos: New default conversation mode. Find photos by having a conversation.
⏳ Progress Tracking: Know when the photos database or AI models are expected to be available.
🖼️ Less Memory Usage: Using the smallest AI models available.
🛠️ Error Reporting: If something goes awry, you can report it.
1 Like
Comment
Included in this version:
📸 Describe Photos: By selecting a photo in the Mac Photos app, the app will display and describe it using local AI models, ensuring your data remains private.
A lot of technical hurdles had to be overcome only to provide this very simple feature... but the tech seems solid - it works!
1 Like
Comment
Updated the website to be less buggy, include some minimal about contents and sharing metadata. I also took some time to think about how Remember This can become an actual product that provides real value to people. It will take a lot of product engineering to get to that state, and as a side project and busy parent I have very limited time to work on it, but it will be fun to work on it, it has to be done. Now, focus can shift to assembling a first technical preview of the Desktop app.
1 Like
Comment
After researching different components for the perfect architecture for the first version of the app, I settled on a combination of:
- Ollama for Local LLM
- osxphotos for Apple Photos metadata extraction
- Electron for Desktop app packaging
- Open WebUI (based on Sveltekit) for the web-based UI
- Python and FastAPI for the backend API
- Redislite for tracking real time state updates between backend and frontend
- Dagster to run local ETL jobs related to photo collections and LLM querying
This should allow the app to do what I envision it could eventually do: visualize your kids' life histories - targeting busy parents that want to be able to look back and remember the early years and feel proud about how their kids are developing (coincidence that I match the target customer profile 100%? :) )
1 Like
Comment
I managed to settle on a product name (Remember This), buy a domain, learn Svelte and scrape together a minimal website at https://rememberthis.ai. There is almost no content, mobile styles are a bit janky and untested, but it's published! Having a website makes the whole project more real and motivating to work on.
1 Like
Comment
After documenting and scoring 104 product ideas (a process that happened over two months), I shortlisted 6 of them and began a shotgun approach: Prototyping all of them in parallel.
By April 27, I eventually chose to focus on an app for personal memory exploration using local AI. Because of the incredible business opportunity? No, simply because it was the idea that I felt most passionate about, an idea that I could see myself working on for as long as it takes to make it into something great.
1 Like
Comment
About
Memories are, in the end, all we got. I and many others take thousands of photos each year but get little value of them. I want to help people use them to connect with friends and family

Comment