Overview

‍

The knowledge base is the structured memory behind everything Sera writes. It grows from your interviews, your uploads, and your connected sources — meeting recordings, forwarded emails, documents — and it is the reason Sera's drafts sound like you and cite what you actually know, rather than generic filler.

‍

‍

Three ways to browse

‍

• Spine: Your knowledge organized by topic — the top-down view of what Sera knows.
• Graph: Entities and relationships — the people, companies, and topics in your world and how they connect.
• Manage: The working view for reviewing and maintaining what is stored.

‍

‍

Entity extraction

‍

As material flows in, Sera extracts entities — people, companies, and topics — and keeps them linked. Mention a customer in an interview in March and again in a forwarded email in June, and the knowledge base knows they are the same customer.

‍

‍

Freshness reconciliation

‍

Knowledge goes stale: positioning changes, numbers get updated, companies rebrand. Freshness reconciliation keeps stale facts from resurfacing in drafts, so last year's pitch does not leak into this quarter's posts.

‍

‍

Yours, in markdown

‍

Every node of the knowledge base is markdown — readable, portable, and exportable at any time, including after you cancel. It is never used to train a model that serves another customer. You are building an asset you own, not feeding someone else's model.

‍

‍

How it fills up

‍

1. Weekly interviews and daily SMS check-ins.
2. Discovery sections and re-entries.
3. Direct uploads and the Google Drive picker.
4. Email forwarded to your Sera inbound address.
5. Fathom meeting recordings.