LiftCast: A Local-First AI Workout Forecaster Built for a Friend

Description and use

LiftCast is a local‑first AI workout logger that turns informal chat notes into structured data, forecasts progress, detects plateaus, and provides a 10‑second audio recap, all offline.

Problem

Armaan never logs workouts, so he cannot assess strength progress or stalls.

Outcome

LiftCast provides structured logging, progress forecasting, plateau detection, and audio briefing entirely offline, enabling Armaan to track strength progress without manual entry.

Workflow

• Paste WhatsApp note into 10‑second shorthand logger
• Gemma parses note into JSON schema
• Data stored locally
• TabPFN forecasts next session top set with 95% CI
• 56‑day regression flags stalls
• Barbell visualizer shows plate layout
• ElevenLabs synthesizes 10‑second audio recap
• Play recap in gym earbuds

Evidence

LiftCast is a local‑first tracker and strength forecaster built for his workflow

Source

10‑Second Shorthand Logger: Pastes raw text, typos, or Hinglish notes. A local Gemma model parses them into structured sets via an enforced JSON schema.

Source

In‑Context Progress Forecaster: Prior Labs' TabPFN runs locally on CPU to forecast his next session's top set with calibrated 95% prediction intervals.

Source

Plateau Detection & Barbell Visualizer: Uses a 56‑day regression slope to flag genuine stalls (ignoring routine deloads) and renders a color‑coded barbell sleeve graphic showing exact plates to load.

Source

Hands‑Free Audio Briefing: Synthesizes a 10‑second voice recap via ElevenLabs (with offline browser speech fallback) straight into his gym earbuds.

Source

Data Sovereignty: Workout logs, bodyweight, and personal notes stay in a local SQLite database (data/liftcast.db). Zero cloud egress.

Source