General
Do I need to be a developer? No. The no-terminal path is three steps: connect Strava, upload the skill zip in Claude, ask “how am I today?”. No code, no keys.
Do I need Strava? No. Strava gives feezify the objective half (your training load). Without it, feezify still works from your journal alone — you just won’t get a TSB. You can add Strava later anytime.
Is it free? Yes — free and open source (Apache-2.0), built in public. If it helps you, you can support the work on Tipeee → https://www.tipeee.com/byNicolasJD.
Will it tell me what workout to do? No, by design. feezify is a copilot, not a coach: it reads your day (green/amber/red and why) and you decide. It never prescribes workouts, paces, watts, or durations, and never gives medical advice. See what is feezify?.
Does it send my data anywhere? No. It’s local-first — no server, no account, no telemetry. It only touches a third party (Strava) after you explicitly consent, and that’s revocable. See privacy & consent.
No-terminal path (Claude)
The skill won’t run / nothing happens when I ask. Make sure code execution is enabled for the skill (Settings → Capabilities → Skills). The skill runs a tiny bundled Node engine to compute your numbers.
Where do I get the skill zip?
From the latest release —
feezify-skill-<version>.zip. Upload it in Claude: Settings → Capabilities → Skills →
Upload skill.
My journal/memory disappears between chats.
A plain web chat has no persistent files. Use Claude Cowork and give it a folder you
own (e.g. Documents/feezify) so your core persists. The copilot will warn you when it’s
in a session-only context and can hand you a zip to keep.
The Strava connector isn’t available to me. The official connector is currently limited to Strava subscribers. If you can’t use it, either skip Strava (journal-only mode works) or use the developer path with your own API tokens.
Developer path
git check-ignore .env prints nothing.
Then .env is not ignored — stop and fix your .gitignore before adding any real
token. It must print .env. Never commit a token.
Strava isn’t being read even though I set a token.
Strava is queried only when both the token is in .env and
connectors.strava: true in your config.yml. Check the consent flag — it’s false by
default. See privacy & consent.
The read shows a form number but the score looks off / neutral.
If you wrote no journal entry for that date, the readiness score is a neutral default
(feezify says so plainly). Add a journal/YYYY-MM-DD.md entry for a real read.
How do I run a read manually?
node --env-file=.env dist/lecture.js ~/my-feezify-core 2026-06-29
(Omit --env-file=.env if you have no .env yet.) It prints JSON: light, tsb +
tsbProvenance, score + scoreProvenance, reason, vigilance, and
recentNarrative.
How do I rebuild the distributable skill zip?
pnpm package:skill
Output: packaging/feezify-skill-<version>.zip (git-ignored build artifact).
Tests / typecheck?
pnpm test # unit tests (the domain math is fully covered)
pnpm typecheck # tsc --noEmit
Understanding the read
What do green / amber / red mean? Green = go; amber = caution (likely hidden fatigue); red = back off (always red for injury or illness). Full rule in the method.
What’s TSB?
Your form — fitness minus fatigue (CTL − ATL). Bands and details in
the method.
Why did feeling override the numbers? Because that’s the core rule: the subjective gates the objective. Low readiness is never green; injury/illness is always red — no number overrides how you actually feel.
Still stuck?
Open an issue on the repo with what you expected and what you saw — feedback is welcome
(this is an early beta). See author.md for who’s behind it.
Back to → Docs index