Skip to content
View RonanDavalan's full-sized avatar

Block or report RonanDavalan

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. 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
RonanDavalan/README.md

Ronan Davalan

Systems administrator and self-taught developer, independent. Debian and Linux infrastructure since 1997, supervision of AI models since 2023.

I build reliable, lasting solutions for small businesses and independents: automation scripts, sovereign local networks, command-line tools.

What my tools do

Voice and accessibility

  • vosk-cli-dictation: real-time dictation for Linux, offline, from the command line.
  • PiperRead: reads aloud any selected or copied text with a neural voice, no cloud service.
  • WhispSkrid: push-to-talk dictation for Linux: hold a key, speak, release; offline transcription typed at the cursor.

Tools for language models

  • mcp-server-diwall-browser: an MCP server that gives a model the screenshot, the accessibility tree and the gestures of the browser where the operator is signed in, under their confirmation.
  • LibreModular-MCP: a universal MCP server in Rust: a model drives GIMP, Krita, Blender, Inkscape and LibreOffice by running one complete script per call, through a Hub that exposes two tools.
  • Diwall: lets a model see a web interface and replay browser scenarios.
  • Dinoer: web research and structured extraction for language-model agents, with sourced reports.

Server security

  • Argus: counts the attacks on Debian and Ubuntu servers (SSH, Fail2Ban jails, CrowdSec scenarios) and keeps the history as charts, with no permanent service and no database.

Windows

  • WindowsOrchestrator: PowerShell scripts for a reproducible configuration of dedicated Windows workstations.

How the work is done

I design and direct; language models write the code under my supervision: Claude Code (Anthropic) on most projects, Google Gemini, Grok and Claude on WindowsOrchestrator. The credits of each project say who did what; for WindowsOrchestrator, they are in section 8.4 of its developer guide.

Elsewhere

The project websites are available in English, French, German and Spanish.

En français : parcours, réalisations et CV sur davalan.fr, qui est en français ; les sites des projets existent aussi en français.

Pinned Loading

  1. vosk-cli-dictation vosk-cli-dictation Public

    Offline voice dictation for Linux that types recognized speech into any application: Vosk engine, tray icon, optional Whisper refinement, native packages for Debian, Fedora, openSUSE and Arch.

    Python 10 2

  2. PiperRead PiperRead Public

    Privacy-First Neural Text-to-Speech for Linux (Wayland & X11).

    Python 3

  3. diwall diwall Public

    Execution API that gives language models (Claude Code, Gemini…) the ability to see web interfaces — Playwright · Set-of-Mark · ReAct

    Python

  4. mcp-server-diwall-browser mcp-server-diwall-browser Public

    Diwall MCP — the operator's browser, seen and handled by a language model. An MCP server that gives the model the screenshot, the accessibility tree and the gestures of the browser where the operat…

    Python

  5. whispskrid whispskrid Public

    Push-to-talk dictation for Linux — hold a key, speak, release; offline Whisper transcription typed at the cursor.

    Python

  6. WindowsOrchestrator WindowsOrchestrator Public

    Your autopilot for dedicated Windows workstations. Configure once, and let your system manage itself autonomously.

    PowerShell 1