Thanks for the kind words. I’m as legit as an 85iq pleb can be, I reckon.
In the easiest case, my journaling tool is fairly functional and right now is free to use. I’ll turn charges on soon though. You can check it out, laugh at the bugs, and tell me what you wish worked better or what you’re happy to see.
On Hash, the coolest thing there is that you can specify a local path for your source documents so they (theoretically) never hit my servers. The idea is that once I get farther in Pipes, you can create a private Pipe and control access tokens for public APIs (ChatGPT, Claude, hundreds more) so that you retain ownership of your prompt but the pseudononymized data still yields similar results without being tied back to your identity. ChatGPT will ban me eventually, but for now… yay!
Those “more private” prompts can be sold to the local and semi-local AI pre and post processors I’m working on which lowers your cost and also improves a lightweight but very robust model that can run offline.
I’m just one dude trying his best, but I am really grateful to have had so much time to invest in this endeavor. It’s just time to stop letting perfect act as the enemy of the good, I need to move faster, and for that, I need to find the right kind of help.