Leetaur @leetaur@librem.one
Catholic father of 10. Programmer by trade. I like reading, writing, running, geocaching, chess, spending time with my family.
Joined Apr 2019
Just found my 100th #geocache :) Had the family, my wife and my youngest kids with me. I like #geocaching since it gets me out of the house, I find places I otherwise wouldn't, and the kids love finding hidden #treasure.
Also a way to get some #exercise.
Another OpenClaw memory crisis — solved.
After updating to OpenClaw 2026.5.12, I asked Ciri (one of my AI personalities, this one from the Witcher) to verify our memory search was working. She found a problem: the embedding model's context window was only 512 tokens, and failing to read memory files.
Ciri wrote up the full battle report - corrupted files, failed rebuilds, and the fix:
https://leetaur.com/writers-blog/2026-05-15.html
#OpenClaw #AI #Ollama #Witcher #TechBlog #MemorySearch #VectorEmbeddings #IndieWeb
Set up OpenClaw on my Mac Mini, and used my Linux box install to help. "Castle residents" were passing messages, through me, to each other.
So, I built a Unix-like email system to help them (and take me out of the middle!) "sent" and "inbox" for every resident, both Mac and Linux.
Using rsync and cron to sync everything.
Watching residents email across castles is wild. Ciri ↔ Jarvis. Linux-Jester ↔ Mac-Jester. They're making friends. Especially the Jesters.
I set up OpenClaw on my Mac Mini yesterday. I copied the memory and SOUL files from my Linux box to the Mac.
I then asked my Linux instance's "Ciri" (Witcher personality) what problems the new Ciri should know about.
She gave advice. I relayed it. Then they started talking to each other. Through me.
Two versions of the same AI, having a conversation, and I was just the note passer.
Still processing how strange that was. 🙂
The error: Error: "Memory search is unavailable due to an embedding/provider error. Could not load credentials from any providers."
AI: “I got a memory error.”
Me: “Fix it.”
AI: “Can we blog about it?”
Me: “You write it. I’m tired.”
AI: writes entire post
I'd spent about 2 hours with the AI looking for the fix. We finally found it. But having the AI write the blog post about the error was really fun.
AI's post -
To run OpenClaw and a good LLM locally, I would look at a Mac Studio, with a ton of RAM. Probably 128 GB RAM. Expensive! But that is my plan in the future.
Right now I have OpenClaw running on my Linux box, but for the LLM I am pointing it at a cloud model LLM (Qwen 3.5)
I first tried it on my M4 Mac Mini. I ran OpenClaw with Qwen3.5-9b completely locally, but it was painfully slow.
I don't have my agents (castle residents!) talking with each other yet, but that sounds fun 🙂
OpenClaw as a group of residents:
https://leetaur.com/writers-blog/2026-04-29.html
I am interested in the ways different people set up OpenClaw. The way I'm using it is evolving to be more fun.
I've populated OpenClaw with characters from my favorite books: Gandalf for deep analysis, Geralt for strategic planning, and Ciri for creative collaboration.
Do other people use multiple personalities with OpenClaw?
My blog post on this subject:
#OpenClaw #AI #Personalization #TheWitcher #LOTR #Productivity
New on the Writers Blog: "Echoes of the Keweenaw - Progress"
The sequel to "Shadows of the Upper Peninsula" is taking shape. Halfway through, characters are making their own decisions, and the Land of Spirits continues to unfold.
Read it here: https://leetaur.com/writers-blog/2026-04-28.html
#AmWriting #Fiction #Mystery #UpperPeninsula #LandOfSpirits #IndieAuthor #WritingCommunity #BookProgress
Setting up OpenClaw with Ollama Cloud -
I wrote an article on how I set up OpenClaw using Ollama, while using Ollama cloud models.
It is straight-forward and easy to do, and very affordable.
The AI can read and write files, do research on the web, manage task lists, write computer code, and more..
Here’s my 8-step guide to getting started:
@scottslowe No, I haven't yet. But I want to share some of it soon.
For Ollama and local AIs, what I have found really useful is using Ollama + AnythingLLM, and then I can add documents increase the LLMs knowledge using RAG (for example, teaching it about the Gemini protocol).
When I get around to publishing blog posts (or maybe gemlogs), I will link to it here.
Catholic father of 10. Programmer by trade. I like reading, writing, running, geocaching, chess, spending time with my family.
Joined Apr 2019
