For V1 Customers

    V1 to V2
    Migration

    Move your markdown brain to V2's structured layer in 30 minutes. Your data, your backend, your call.

    What Changed in V2

    • Structured memory layer. V1 stored everything in memory/*.md. V2 adds a SQL database (PocketBase local OR Supabase BYOD) that the second-brain-data MCP reads and writes.
    • Progressive disclosure. CLAUDE.md is shorter. Domain rules auto-load from .claude/rules/ only when you touch matching files. Less noise per turn.
    • Audit-resilient by default. The decisions table captures every choice with rationale and rejected alternatives. Queryable, exportable, regulator-friendly.
    • Three pre-wired MCPs. Your V2 repo ships with second-brain-data, second-brain-guide, and second-brain-health-check already configured with your SBF_TOKEN.

    The Migration in 4 Steps

    1. 1.
      Receive your V2 repo
      If you bought a V2 package and never received the repo, the questionnaire link is in your purchase email. Fill it in and the repo gets generated within 24h with your SBF_TOKEN baked into .claude/settings.json.
    2. 2.
      Pick your database backend
      PocketBase (local, default) or Supabase (cloud, BYOD). Both run through the same MCP. See the access page for the trade-off table.
    3. 3.
      Drop your V1 markdown into V2's memory/ tree
      Place files in subdirectories by category: semantic/, company/, customers/, workflows/, etc.
    4. 4.
      Run /migrate-knowledge
      The skill auto-types files based on the directory (semantic → pattern, company → fact, workflows → reference) and writes them into your DB. Idempotent. Run twice without duplicating.
    DON'T PICK THE WRONG SKILL

    /migrate-knowledge ≠ /import-memories

    Your V2 repo ships with two similarly-named skills. They do completely different things.

    /migrate-knowledge
    V1 markdown + cloud knowledge → your local SQL DB. This is the V1 → V2 path.
    /import-memories
    ChatGPT/Claude conversation exports (conversations.json) → memory/imported/. Different use case. Don't confuse them.

    "I work across iPhone, iPad, and a NAS — does V2 break that?"

    No, but you should pick the backend on purpose, not by default. Both work; they trade differently.

    PocketBase

    Single binary. Run on your laptop or NAS (Synology/QNAP, ARM-friendly). Point clients at SB_POCKETBASE_URL. Full local sovereignty.

    Supabase

    BYOD cloud. True cross-device with no NAS dependency. Set SB_BACKEND=supabase + URL + service key. Schema auto-spins on first MCP call.

    Markdown still git-syncs alongside the DB. Hybrid is fine — keep memory/ for offline iOS reading, use the DB for queryable structured recall.

    Get the Playbook

    Full migration walkthrough with copy-paste commands.

    • 4-step migration walkthrough
    • Backend chooser (PocketBase vs Supabase)
    • Combining V2 + GTM repos
    • Compliance/audit considerations

    No spam ever.

    Iwo Szapar
    Created By

    Iwo Szapar

    • •Shipped both V1 and V2 Second Brain — and migrated his own data twice
    • •Co-created AI Maturity Index with Harvard
    • •Trained 25,000+ professionals on AI

    Get the Migration Playbook

    Free. Step-by-step commands. Both backends covered.

    Instant access. No spam ever.

    Don't Have V2 Yet?

    Second Brain 2.0

    Pre-configured Claude Code repo. Structured memory layer, 30+ agents, three pre-wired MCPs.

    DIY $197 · Kickstart $597 · DWY $2,497

    V1 vs V2
    MemoryOS

    The paid MCP tools tier. 40 tools that read/write your local DB. Kickstart includes Pro for first year.

    Free · Standard $199/yr · Pro $349/yr

    Get MemoryOS