MemoFSMemoFS

How to ingest Notion Pages into MemoFS

Ingest Notion workspace pages and databases into structured MemoFS markdown memory files.

This cookbook shows you how to connect your team's Notion workspace or project database to MemoFS, enabling AI agents to recall architecture specifications, product requirements, and knowledge docs written in Notion.

Prerequisites

  • Node.js: >= 22.0.0
  • Notion Internal Integration Token (ntn_... or secret_...)
  • Notion Database ID or Page ID shared with your integration.

Step 1: Add Notion Secret

Store your Notion integration token in .memofs/secrets.json:

.memofs/secrets.json
{
  "notion_token": "ntn_yourNotionIntegrationToken"
}

Ensure .memofs/secrets.json is added to your .gitignore so secrets are never committed to version control.

Step 2: Configure the Notion Connector

Register the Notion connector using the CLI. You can specify a databaseId or a searchQuery:

npx @memofs/cli connectors add \
  --id team-notion \
  --type notion \
  --secret-ref notion_token \
  --source-mapping '{"databaseId":"32characterDatabaseIdHere","limit":50}'

Or for full-workspace keyword search ingestion:

npx @memofs/cli connectors add \
  --id team-notion \
  --type notion \
  --secret-ref notion_token \
  --source-mapping '{"searchQuery":"Architecture Spec","limit":25}'

Step 3: Run Ingestion

Ingest pages into local memory:

npx @memofs/cli connectors run --type notion

Output:

✓ Connectors run complete (25 written, 0 skipped)
Connectors run summary:
- ran: team-notion
- written: 25
- skipped (already ingested): 0

Step 4: Verify Recall

Test retrieval of the Notion documentation:

npx @memofs/cli context --query "What is the data retention policy for user uploads?"

On this page