Skip to content
View Efs-O's full-sized avatar
🎯
Focusing
🎯
Focusing
1 follower · 2 following

Block or report Efs-O

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Efs-O/README.md

Efstathios Outas · evolvlabs

I build tools for running AI models on your own GPUs, and I work on making open models better at Greek. Based in Greece. More at evolvlabs.dev.

Main projects

Project What it is
Forge VS Code coding agent for your own models: full llama.cpp/GGUF control, 60+ agent tools, undo any turn, no telemetry. On the VS Code Marketplace and Open VSX.
HalluScribe Local-first archive and search for your AI coding sessions (Rust + Tauri), with an MCP server so agents can recall earlier sessions.
NemotronGR Case study: a Greek LoRA fine-tune of NVIDIA Nemotron 3 Nano 30B-A3B on 44 GB of consumer GPUs, natively on Windows. Weights on Hugging Face.
Gemma4GR Greek fine-tune of Gemma 4 E4B, and JOY, an open Greek Piper TTS voice.
Gemma4kids Offline coding teacher for kids, powered by Gemma 4: plain English in, HTML animations out.

Smaller tools

  • CacheWarden: keeps Claude Code's prompt cache warm.
  • forge-relay: a coordination board for Claude Code and Codex agents working in one repo.
  • HalluMeter: a desktop ring that warns when an AI session's context window is filling up.
  • airos-wakebot: wake a PC from Telegram, with the bot running on a Ubiquiti radio.

Everything is built and tested on a Windows workstation with 2× RTX 5060 Ti and an RTX 3060.

Contact: hello@evolvlabs.dev

Pinned Loading

  1. Forge Forge Public

    Run your own models in VS Code: full llama.cpp/GGUF control, 60+ agent tools, undo any turn, no telemetry.

    TypeScript

  2. halluscribe halluscribe Public

    Local-first AI coding-session archive and search app

    Rust

  3. Hallumeter Hallumeter Public

    Always-on-top desktop ring that warns when your AI coding session's context window is getting full. Claude Code, Codex, Cursor, Continue, Forge. Tauri + React.

    Rust

  4. CacheWarden CacheWarden Public

    VS Code extension that keeps Claude Code's Anthropic prompt cache warm — auto-pings before the 5-minute TTL expires.

    TypeScript 4

  5. Gemma4GR Gemma4GR Public

    Greek fine-tune of Gemma 4 E4B (v3: LoRA merged and verified tensor by tensor), plus JOY, an open Greek Piper TTS voice.

    Python 2

  6. NemotronGR NemotronGR Public

    Case study: Greek LoRA fine-tune of NVIDIA Nemotron 3 Nano 30B-A3B on 44 GB of consumer GPUs

    Python