dubs

A proof ledger. Nothing more.

dubs stores immutable, first-order facts about testimonials, praise, or public statements. It is an append-only record of truth with provenance.

• Not a marketing tool
• Not a funnel system
• Not a testimonial gallery
• Not a CRM
• Not a rendering engine
• Not a permissions system

Store facts, never interpretation.

The ledger captures exactly what was said, who said it, where it came from, and when it was captured. That's it.

Anything derived from these facts—summaries, claims, permissions, views, audiences, formats, marketing usage—lives OUTSIDE the core data. The schema is the immutable core.

dubs.json is append-only. Past entries are never edited. If a correction is needed, add a new entry.

{
  "version": 1,
  "dubs": [ ... ]
}

Each Dub record contains:

• id — UUID
• who — The person (name, social profile, url)
• original — The statement (text, source url, optional media blob)
• event — Context (name, date, url)
• captured_at — When this fact was recorded

See schema.json for the exact shape.

Filter by person:

jq '.dubs[] | select(.who.name == "Matthew Hall")' dubs.json

Filter by event:

jq '.dubs[] | select(.event.name == "AI Bootcamp")' dubs.json

Filter by date range:

jq '.dubs[] | select(.event.date >= "2024-01-01" and .event.date <= "2024-12-31")' dubs.json

Find entries with media:

jq '.dubs[] | select(.original.blob != null)' dubs.json

Count all entries:

jq '.dubs | length' dubs.json

The ledger will eventually move to a database (e.g., Supabase) for better querying and scale. The schema will remain identical. dubs.json serves as the canonical source until migration.

Do not mutate historical entries.

If you find an error in an existing record, do not edit it. Add a new entry with the correction. The ledger is append-only by design.