Why Neotoma exists
Personal data stays fragmented across email, downloads, chats, cloud drives, exports, and AI threads. Provider memory stays conversation-shaped and platform-locked, so it cannot unify documents, resolve entities across sources, or rebuild timelines with deterministic provenance.
Neotoma is the state layer for structured personal data: schema-first writes, hash-based entity IDs, immutable observations, explicit provenance, and cross-tool access via MCP and the HTTP API. Agents and apps share one substrate instead of each reinventing extraction, merges, and audit trails.
The canonical long-form narrative lives in the repository under docs/foundation/problem_statement.md (Neotoma checkout). This page is the short public surface aligned with that doc.
Next: Architecture, Memory guarantees, Walkthrough.