I came up with the idea to combine two of the most effective learning approaches, Anki Cards and visual Memory Palaces, in order to created an application for automatised creation of visual scenes based on selected Anki content.
The result is my application Cranky - the memory palace generator for anki cards
Users can select up to 25 cards from one Anki deck or from a tag selector, based on difficulty, repetition count or random selection.
The card content is then processed and uploaded to a GPT endpoint to generate a visual scenery based on a theme selected by the user.
Currently the client logic lives as a local python script and interacts via Anki connect.
Basically Anki is developed by volunteers so self-promotion of projects like freemium is not welcomed by the average Anki users (In other words funding like general app stores or Kickstarter is very rarely successful), so I recommend that the server code be open sourced and made available to advanced users for their own use if possible. (e.g. popular add-ons AnkiBrain and HyperTTS offer paid plans, but advanced users can register their own API keys and use add-ons, and the code is open source.)
The average Anki user, being the target audience of the current project, cannot be expected to have an OpenAI as well as Replicate API key under his disposal and simultaneously run a server for rendering the Cranky web interface.
I was interested to create an application that an average Anki learner can use, with an existing and functional infrastructure behind it and which is free of charge.
There is no obligation to open-source the complete server logic from the very beginning.
Your argument about “freemium” not being welcomed does also not really apply here an in this ecosphere. Despite the fact that I myself created several free Anki Add-ons in the past.
At the same time, I happily paid for the Anki IOS app (which cross-finances the app development).
In addition, I also have a paid annual subscription to AnkiHub.
Furthermore, I paid a 3 digit amount to authors of commercial Anki Decks.
In this regard, I am very happy to financially support projects, which develop useful learning tools and developers who deliver quality content.
Also, the current setup is not yet set in stone.
For the time being, I would like to collect some impressions from actual users, improve the prompt logic and user interface.
Then I can decide about the next steps. Putting the complete stack up on Github can still be an option for the future.
