Private tool · Fictional public demo

A relationship memory, not a sales pipeline.

My Personal CRM resolves scattered contact identities into one person, remembers why the relationship matters, and suggests who may be worth contacting. AI can research missing context, but I approve every fact.

The real database and contact data remain private

A sanitized Personal CRM screen populated with fictional records.
Fictional records used for the public preview
1canonical record per person
Manyemails, handles, and source identities
100%provenance on imported facts
Reviewrequired before enrichment writes

Faithful public edition

Use the real interface with fictional records.

The navigation, views, scoring panels, and review queues mirror the private product.

Personal CRM · public demo

The demo uses the production interface with invented people and organizations. Actions last only until you reload the page.

The architecture

Identity first. Provenance on every claim.

LinkedIn, WhatsApp, email, and manual notes all describe people in different ways. The CRM keeps those source identities separate and links them to one canonical person only when the evidence supports it.

  • Importers retain the source, batch, date, and confidence behind each field.
  • Relationship scores combine importance, recency, and interaction evidence.
  • AI enrichment produces a proposal. Accepting it writes the sourced facts; rejecting it leaves the database unchanged.

Product judgment

Private data changes what counts as a good feature.

Raw message archives would add context, but they would also create a much larger privacy and security burden. The system stores structured interaction metadata and curated notes instead.

The production interface stays private. This public version shows the interaction model with invented records rather than redacted personal data.