Meridiona/meridian

Stop letting your work go unnoticed. Meridian keeps an automatic work journal of what you worked on and what you finished, summarised daily, and automatically drafts your project tickets for you.

Meridian – Automatic Day‑Reconstruction & Work‑Log Assistant

What it is – Meridian is a desktop application (macOS Apple Silicon or Windows 10/11) that continuously records what you do on your screen, stores the data locally in an encrypted database, and uses an LLM (via any AI provider you connect) to generate useful summaries: a timeline of the day, an end‑of‑day stand‑up report, and draft work‑log updates that can be pasted into Jira.

Why it matters – Many knowledge workers spend time manually recalling or writing up what they did each day. Meridian automates that by turning raw screen activity into structured information, helping you:

  • Remember exactly what you worked on, even weeks later.
  • Produce stand‑up notes and ticket updates without typing them.
  • Analyse where your time actually went, which can surface estimation errors.

Key features (as described in the README)

  • Continuous screen capture – runs as a low‑overhead background daemon (5‑10 % CPU, 0.5‑3 GB RAM) and stores about 20 GB of data per month (pruned after 30 days).
  • Day reconstruction – a visual timeline you can scrub through like a video.
  • End‑of‑day summary – automatic list of completed tasks, unexpected fixes, and a ready‑to‑paste stand‑up.
  • Work‑log drafting – generates Jira‑compatible updates that you only need to review and send.
  • Privacy‑first – all raw activity stays on‑device in an encrypted DB; only the minimal text needed for a summary is sent to the LLM you configure. Diagnostic data is opt‑out and anonymised.

Technology stack

  • Rust – core daemon that captures screen activity and manages the encrypted database.
  • Tauri – provides the cross‑platform desktop UI (tray icon, windows).
  • Node 20+ / Bun – used for the front‑end tooling and build scripts.
  • LLM integration – the app does not ship its own model; you connect any provider (e.g., OpenAI, Anthropic) and the app sends only the summary text.

Getting started

# Clone and set up the repo
git clone https://github.com/Meridiona/meridian
cd meridian
cp .env.example .env

# Build everything (Rust, Node/Bun, Tauri)
bash install-dev.sh

# Install git hooks (run once)
bash scripts/setup-hooks.sh

# Run the app in development mode (opens two terminals)
bash dev-start.sh

The script launches the Rust daemon and the Tauri tray; both rebuild automatically when you edit source files.

Installation for end users – The README links a download button that points to https://meridiona.com/dev?ref=github-readme#download. The binary is packaged for the supported platforms and includes the background daemon and tray UI.

License & usage

  • The code is source‑available under the PolyForm Strict License 1.0.0 (personal, non‑commercial use only). Modification or redistribution is prohibited without a separate commercial license. Earlier releases remain under MIT.

Who might benefit

  • Software engineers, product managers, or anyone who logs work in Jira or similar tools and wants to reduce manual reporting.
  • Teams that need accurate time‑tracking for retrospectives without installing heavyweight time‑tracking software.

Limitations

  • Only macOS Apple Silicon (Intel not supported) and Windows 10/11 are officially supported.
  • Data is kept for 30 days; older history is automatically pruned.
  • Commercial use requires a paid license from Meridiona.

All information above is taken directly from the repository’s README; no additional features have been inferred.

Related

  • Project
  • Project
  • Project
  • Project
  • Project