memory-schema

Memory Schema

Manage structured note types using Basic Memory's Picoschema system. Schemas define what fields a note type should have, making notes uniform, queryable, and validatable.

When to Use

• New note type emerging — you notice several notes share the same structure (meetings, people, decisions)
• Validation check — confirm existing notes conform to their schema
• Schema drift — detect fields that notes use but the schema doesn't define (or vice versa)
• Schema evolution — add/remove/change fields as requirements evolve
• On demand — user asks to create, check, or manage schemas

Picoschema Syntax Reference

Schemas are defined in YAML frontmatter using Picoschema — a compact notation for describing note structure.

Basic Types

schema:
  name: string, person's full name
  age: integer, age in years
  score: number, floating-point rating
  active: boolean, whether currently active

Supported types: string, integer, number, boolean.

Optional Fields

Append ? to the field name:

schema:
  title: string, required field
  subtitle?: string, optional field

Enums

Use (enum) with a list of allowed values:

schema:
  status(enum, current state): [active, blocked, done, abandoned]

Optional enum:

schema:
  priority?(enum, task priority): [low, medium, high, critical]

Arrays

Use (array) for list fields:

schema:
  tags(array): string, categorization labels
  steps?(array): string, ordered steps to complete

Relations

Reference other entity types directly:

schema:
  parent_task?: Task, parent task if this is a subtask
  attendees?(array): Person, people who attended

Relations create edges in the knowledge graph, linking notes together.

Validation Settings

settings:
  validation: warn    # warn (log issues) or strict (errors)

Use strict as the canonical enforcing mode. error is accepted only as a compatibility alias.

Complete Example

---
title: Meeting
type: schema
entity: Meeting
version: 1
schema:
  topic: string, what was discussed
  date: string, when it happened (YYYY-MM-DD)
  attendees?(array): Person, who attended
  decisions?(array): string, decisions made
  action_items?(array): string, follow-up tasks
  status?(enum, meeting state): [scheduled, completed, cancelled]
settings:
  validation: warn
---

Discovering Unschemaed Notes

Look for clusters of notes that share structure but have no schema:

1. Search by type: bm_search(note_types=["meeting"]) — if many notes share a type but no schema/Meeting.md exists, it's a candidate.

2. Infer a schema: Use bm_schema(command="infer") to analyze existing notes and generate a suggested schema:

   bm_schema(command="infer", note_type="Meeting")
   bm_schema(command="infer", note_type="Meeting", threshold=0.5)  # fields in 50%+ of notes
   

   The threshold (0.0–1.0) controls how common a field must be to be included. Default is usually fine; lower it to catch rarer fields.

3. Review the suggestion — the inferred schema shows field names, types, and frequency. Decide which fields to keep, make optional, or drop.

Creating a Schema

Write the schema note to schema/<EntityName>:

bm_write(
  title="Meeting",
  directory="schema",
  note_type="schema",
  metadata={
    "entity": "Meeting",
    "version": 1,
    "schema": {
      "topic": "string, what was discussed",
      "date": "string, when it happened",
      "attendees?(array)": "Person, who attended",
      "decisions?(array)": "string, decisions made"
    },
    "settings": {"validation": "warn"}
  },
  content="""# Meeting

Schema for meeting notes.

## Observations
- [convention] Meeting notes live in memory/meetings/ or as daily entries
- [convention] Always include date and topic
- [convention] Action items should become tasks when complex"""
)

Key Principles

• Schema notes live in schema/ — one note per entity type
• note_type="schema" marks it as a schema definition
• entity: Meeting in metadata names the type it applies to
• version: 1 in metadata — increment when making breaking changes
• settings.validation: warn is recommended to start — it logs issues without blocking writes

Validating Notes

Check how well existing notes conform to their schema:

# Validate all notes of a type
bm_schema(command="validate", note_type="Meeting")

# Validate a single note
bm_schema(command="validate", identifier="meetings/2026-02-10-standup")

Important: bm_schema(command="validate") checks for schema fields as observation categories in the note body — e.g., a status field expects - [status] active as an observation. Fields stored only in frontmatter metadata won't satisfy validation. To pass cleanly, include schema fields as both frontmatter values (for metadata search) and observations (for schema validation).

Validation reports missing required fields (as observation categories, or relations for entity-reference fields) and invalid enum values. Undeclared observation categories and relations are listed as informational "unmatched" items. Scalar values are not type-checked.

Handling Validation Results

• warn mode: Review warnings periodically. Fix notes that are clearly wrong; add optional fields to the schema for legitimate new patterns.
• strict mode: Use where conformance matters (e.g., automated pipelines consuming notes).

Detecting Drift

Over time, notes evolve and schemas lag behind. Use bm_schema(command="diff") to find divergence:

bm_schema(command="diff", note_type="Meeting")

Diff reports:

• Fields in notes but not in schema — candidates for adding to the schema (as optional)
• Schema fields rarely used — consider making optional or removing
• Cardinality changes — a field that moved between single-value and array

Schema Evolution

When note structure changes:

1. Run diff to see current state: bm_schema(command="diff", note_type="Meeting")
2. Update the schema note via bm_edit with the metadata parameter. Top-level keys are replaced whole, so pass the complete updated schema map along with the new version:

   bm_edit(
     identifier="schema/Meeting",
     command="set_metadata",
     metadata={
       "version": 2,
       "schema": {
         "topic": "string, what was discussed",
         "date": "string, when it happened",
         "attendees?(array)": "Person, who attended",
         "decisions?(array)": "string, decisions made",
         "location?": "string, where it happened"
       }
     }
   )
   

3. Re-validate to confirm existing notes still pass: bm_schema(command="validate", note_type="Meeting")
4. Fix outliers — update notes that don't conform to the new schema

Evolution Guidelines

• Additive changes (new optional fields) are safe — no version bump needed
• Breaking changes (new required fields, removed fields, type changes) should bump version
• Prefer optional over required — most fields should be optional to start
• Don't over-constrain — schemas should describe common structure, not enforce rigid templates
• Schema as documentation — even if validation is set to warn, the schema serves as living documentation for what notes of that type should contain

Workflow Summary

1. Notice repeated note structure → infer schema (bm_schema(command="infer"))
2. Review + create schema note   → write to schema/ (bm_write)
3. Validate existing notes       → check conformance (bm_schema(command="validate"))
4. Fix outliers                  → edit non-conforming notes (bm_edit)
5. Periodically check drift      → detect divergence (bm_schema(command="diff"))
6. Evolve schema as needed       → update schema note (bm_edit)