Velohud

velohud turns flat cycling videos into pro broadcast-style content by overlaying real ride telemetry: power, heart rate, speed, elevation and GPS maps, synced to the frame. Problem: Cyclists record hours of footage and activities full of data, but combining them takes pro editing skills. Existing tools are desktop-only, complex, or upload your files to a server. Solution: A web editor. Drop a .FIT/.GPX/.TCX (or Intervals.icu activity) plus MP4/MOV clips, auto-sync by recording date, fine-tune on a time/km graph or click-to-position route map, style 15 widget kinds across 19 templates, and render MP4. Everything runs 100% locally in the browser via WebCodecs. Nothing is ever uploaded. Muse AI at the core: template generation from a prompt, one-click Auto design from the video frame, and surgical AI adjust. All through Muse Spark via the Meta Model API, with data-aware gating so the model only suggests widgets your activity can actually feed. Home artwork is generated with Muse Image. Built for my world: I race bikes and built the tool I always wanted