The whole loop, in order

From install prompt to a living taste hierarchy

Four steps. No dashboard account. The video below is the same path the README uses.

Video

Play this if you want the motion version. The stills underneath cover the same ground.

1. Install and answer the setup prompts

nanotaste setup discovers local coding agents, asks which ones to integrate, asks how often to pull sessions, and can take a personal URL, file, or note as the first seed.

First-run setup discovering agents and asking for a harvest schedule — First-run setup: agents, schedule, optional seed, first harvest.

2. The main file points at category files

TASTE.md is an index. It links to seeded files under taste/. Scheduled harvest writes overlays in taste/learned/ instead of silently rewriting the seeds.

TASTE.md
taste/writing.md
taste/code.md
taste/aesthetic.md
taste/product.md
taste/personal.md
taste/brand.md
taste/communication.md
taste/research.md
taste/learned/*.md

3. Seed whatever actually describes you

A personal site, a bio, screenshots, or a folder of notes. The CLI and the local studio both accept URLs, files, images, and pasted text. Images use the caption and filename; there is no vision model.

Local studio seed form with URL, note, and file inputs — Local studio seed box. Separate from this marketing site.

4. Harvest on a schedule

nanotaste harvest pulls opted-in sessions, extracts lexical signals, refreshes overlays, and writes a report. nanotaste schedule --every weekly --install appends a crontab snippet when crontab is available.

Taste hierarchy with category tags and learned overlays — Hierarchy, tags, and rule counts in the local studio.

Open the studio, not this site, to edit taste

nanotaste serve
# http://127.0.0.1:7468

Command flags, scoring, file locations, and the claim lock are in the docs.