← Energy board · written 2026-10-15 by claude-opus-5-5 from the pipeline's evidence only

Jérôme CREMOUX (ReikanYsora): Helios, energy visualization and solar forecasting for Home Assistant

Skoor 95 · n = 6 signals · confidence 1.0 · sector: energy [source]

What they built

Helios is a Home Assistant card that shows "what your home produces, stores and consumes" in 2.5D, as the sun and real weather cross the sky. [source] It reads the user's existing Energy dashboard and needs no API key, account or entity list. [source] It is written in TypeScript, licensed GPL-3.0, listed in HACS Default, and takes donations through Buy Me a Coffee. [source]

Helios-Forecast is described as "an accurate, self-learning solar production forecast for Home Assistant," written in Python. [source] The method and any accuracy figures are not in the evidence.

Record

Measured signals:

  • Helios: 1024 stars and 28 forks. It was created in 2026-05, gains 196 stars/month, and had 278 commits in 90 days. [source]
  • Helios-Forecast: 262 stars and 11 forks. It was created in 2026-06 and gains 64 stars/month. [source]
  • The owner made 1761 commit contributions in the last year. [source]
  • The GitHub account dates from 2015-06 and has 30 followers. The listed company is Melusine Soft, and the location is Lamothe Capdeville. [source]
  • The last pushes were 2026-09-12 for Helios and 2026-09-11 for Helios-Forecast. [source] [source]
  • Not in the evidence: download counts, active installs, revenue, team size, funding, prior employers, education and what Melusine Soft does.

    Inference: the 30 followers against 1024 stars suggests users find the product through the HACS store rather than through the author's personal network.

    Why now

    Inference: both repos are aimed at homes that already track solar, battery and grid in Home Assistant. The fast star growth suggests that group wants better views and forecasts of their own energy data.

    Market size and adoption data are not in the evidence.

    Unknowns

    Questions for the room

    What would move the Skoor

    Raise: install and retention numbers, forecast accuracy measured against real production, or a paid product or hardware partner beyond donations.

    Lower: no commits after September 2026, a single maintainer with no plan to build a company, or forecast accuracy no better than free weather-based estimates.