New tutorial is out! Learn how you can scrape web pages and save them to a local p2p archive. Includes some code but also a guide on how to prompt the local LLM to generate code snippets to save time or if you're unsure how to do something.

#agregore #p2p #web #javascript #ai

Made another lil #agregore app today. This combines the web's SpeechSynthesis API with Agregore's LLM APIs to talk to a spooky LLM with it responding with a voice as well as text.

hyper://816idd9ddxq8asy68sya1y3du3nyipiszcr6tfyq66x47ha3jxuy/speak_ai.html

Make sure to set up Speech Syntehsis on your machine if it doesn't work initially. On linux I had to set up speech-dispatcher and espeak-ng. This should work fully offline, too!

Mozilla #pocket shutting down is exactly why we should be building on top of #p2p data trasfer foundations. Depending on cloud services means it's just a matter of *when* the service will shut down or make your usual flows inaccessible.

The thing is, with tools like #agregore it's way easier to make and distribute small apps that are local to your or your friend's devices.

Tried out the new #bare runtime on Android. Sadly I'm running into issues getting their native modules to load. Apparently there was a fix a few weeks ago but maybe I'm still missing some sort of update in the stack. ๐Ÿคท Hopefully I can get it working and get #hypercore into #agregore mobile before localfirstconf next month.
Made great progress on reimplementing the hypercore protocol handlers in #agregore using the new "bare" runtime which I'll be using to get it into mobile. #p2p
We've got a new release of #agregore with improvements to "reader mode" and a new "View history" page!

K, I'm gonna work on #agregore for a bit. Not sure what to prioritize since there's a lot of TODOs

Feel free to vote on anything you think I should focus on. OR link to the github issue in this thread.

hypercore support on mobile
71.4%
view history page
14.3%
llm based "translate this to English" feature
14.3%
new protocols (e.g. web3:// )
0%
Poll ended at .
Seems I'm on course for doing about 1.5 presentations a month for this year. :P This month is going to be a presentation on local AI in #agregore Next month I'm gonna do one about the fediverse and bsky and looking at how they work / comparing the structures. Hopefully my new graph visualization thing will get far enough that I can embed some in my talks.

There's a new example app out on the website: The LLM Chat.

Use this as a quick way to use the built in #AI features in #agregore or the foundation for your own AI Chat interface!

hyper://agregore.mauve.moe/docs/examples/llm-chat.html

Looks like #Agregore works fine on Windows 11. Now I'm also more prepared to debug issues if they arise :)