SOMI

S.O.M.I. 🕊 We've just added some new features! V0.9 out now Run your own AI This is an A.I. Framework that lets you deploy and control your own A.I. With added capabilities ⚡ Runs on your own hardware 🔒 No cloud/sign-up required 🪪 Privacy first 💃 Personality construct 🧠 Memory storage 🛠️ Digital tools/skills/contracts 💬 Mainframe panel interface 📲 Telegram interface 🔉 Speech capable 🌎 Web searching ✍️ Research 🧑‍💻 Coding 🧸 Daily helper ⭐ If Somi saves you time or inspires you, consider starring the repo

Good morning. Work has resumed on SOMI by one of the primary dev! 1. EDEN update in progress - a civilization rebuilding pack meant to offer advice and guidance with an underlying LLM driver to help restore some functionality to society in the event of global catastrophe e.g. 2. Planning a quantum hook that can allow us to make use of future quantum hardware as well as a narrative point Current dev @Stardew151 - I look forward to completion of this years long hobby project and who knows maybe we would even tokenize at some point in the future to afford stronger hardware, more services, hire additional devs etc.

I just claimed my .agent domain and joined the .agent community! get yours now and help shape the future of autonomous agents agentcommunity.org/join

Good morning fellow Situation Monitors & Decentralization connoisseurs ☕️ Currently in the works: 🔹 Capability Preservation And Resource Adaptation principle so Somi scales down gracefully and scales up aggressively without framework-level weakening 🔹New workstream: Civilization Continuity And Recovery - Hardware Tiers And Survival Mode - Offline Knowledge Pack Architecture - Federated Node Communication Layer - Continuity Domain Packs And Recovery Workflows - Recovery Drill And Blackout Gauntlet 🔹Rigor testing Memory systems and repairs

V0.92 out now! Welcome to the Mineralization Phase where we added substance to the skeleton Much more improvements underway! S.O.M.I. is now competitive in the realm of self hosted A.I.

One year of chugging along ! Framework moggin #MyXAnniversary

This isn't even my final form

The pillars of our framework 1. Free 2. Exceptional user experience 3. Low latency 4. Flawless operations 5. Modular and expandable

Ever since I was a child growing up with ironman I've always wanted my own personal A.I. Assistant and with modern day technologies I am building just that

Accelerate

Run the pre-release on consumer grade hardware with an @ollama model hoster in the backend - use models that fit!

research script testing internally

Homebrewing ☕

Clawd deze nuts - it ain't truly self hosted

We look forward to the new Ollama update for image generation as we'll soon be updating our scripts to match such demand through the ollama LLM Call backend GSomi

We were doing some testing with doctors handwriting and let's just say even Ai has a 'Hard' time at interpreting squigglies in optical character recognition 🍆

New upgrades to the OCR script coming soon - we're focusing on the telegram Ai bot script for now but will expand to local chat for those seeking to stay within compliance e.g. Our lead dev @0xfrostyflakes has to use it to stay within Health Information Act of America Standards

Ollama's new pre-release v0.14.0 adds support for MLX models image generation on Apple chips! We look forward to upgrades and continued expansion of Ollama capabilities which powers our framework's backend Long live self-hosted A.i. Long live privacy data tools! 🦙💙🐦