Vires is a strength-training app from Nous Ergon: log workouts, build routines, find any exercise with hybrid search, and train with an AI coach that plans around your goals. Join the private beta.
Sports & fitnessPython · TypeScriptGitHub @nousergonOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
Vires — strength-training tracker with hybrid exercise search (AGPL-3.0)
Translation pending
Loading the full README.
Read the full README ↗vires acquirit eundo — "it gathers strength as it goes."
A strength-training tracker. Log workouts, build reusable routines, and find exercises with hybrid keyword + semantic search. Mobile-first PWA — installable to your phone's home screen, works like a native app.
This is the tracking + exercise-library substrate for a larger strength-coach vision; the AI coach (multi-week program generation) is intentionally out of scope for this MVP, but the data model is built to receive it.
RDL or describe a movement
("the hamstring curl where your feet are held down") and find it. The same
embedding index detects near-duplicates when you add a custom exercise.bge-small-en-v1.5, 384-d) cosine, fused with Reciprocal Rank Fusion. No API
keys, no GPU. Same embedding index powers near-duplicate exercise detection.Requires uv and Node 20+.
# Backend (Python 3.13)
uv sync
uv run alembic upgrade head # create the schema
uv run python -m api.db.seed # seed the exercise catalog
uv run uvicorn api.main:app --reload # http://127.0.0.1:8000
# Frontend
cd web && npm install && npm run dev # http://127.0.0.1:5173
Run tests + lint:
uv run pytest
uv run ruff check .
AGPL-3.0-only. See LICENSE, NOTICE, and CONTRIBUTING.md.