diff --git a/CHECKPOINT-encryption.md b/CHECKPOINT-encryption.md
index fc3881d..6bba0a7 100644
--- a/CHECKPOINT-encryption.md
+++ b/CHECKPOINT-encryption.md
@@ -1,3 +1,68 @@
+# Checkpoint — NIP-46 session restore on startup (2026-09-24)
+
+## Where things are
+- Project: `/home/avi/Projects/Keynctr`
+- Branch: `master` @ **`0982dad`** ("feat(nip46): restore saved signer
+  sessions on startup/unlock — no fresh scan"). Previous: `73bf17c`
+  (checkpoint), `f53bc56` (sign timeout leash).
+- Working tree: clean for tracked files. Untracked intentionally NOT
+  committed: `COSMIC_THEME.md`, `KeynectrAppIconPossibility02.jpeg`,
+  `deferred/` (stays deferred).
+- Release binary: **rebuilt at `0982dad`**. Frontend untouched — no
+  renderer rebuild needed. Restart Electron before retesting.
+
+## What was completed this session
+1. **Session restore without a fresh scan (`0982dad`)**: the WIP from
+   the Sep-23 session (item 2 of the previous next-steps) is finished
+   and committed. Amber remembers our *client pubkey* as the identity
+   of an approved connection for its whole life, so the client keypair
+   minted at pairing is now persisted in the vault
+   (`Vault::connection_client_keys` — encrypted under the vault key
+   exactly like connection secrets, keyed by the same `VaultRef`, and
+   re-keyed to the identity ref when the handshake resolves it).
+2. **Re-dial path**: `reactivate_saved_sessions()` picks the active
+   profile's live connection (or any other live one), rebuilds the
+   exact wire identity, re-derives the NIP-44 conversation key, and
+   re-sends `connect` — no QR/bunker scan. It is called at `serve()`
+   for unencrypted vaults and after `UnlockVault` for encrypted ones;
+   a restore failure can never block the GUI.
+3. **Cross-account guard**: restored sessions pin `expected_identity`
+   before dialing; `adopt_identity` refuses (never adopts) a signer
+   that answers as a different account — different Amber account,
+   mistyped bunker, or relay spoof all fail closed.
+4. Legacy connection rows without a stored client key are skipped
+   (they need one fresh scan, after which they become restorable too).
+- Verification (all green at `0982dad`): `cargo test` **216 unit + 5
+  e2e passed / 0 failed** — incl. the new
+  `nip46_session_restore_redials_and_refuses_wrong_identity` e2e
+  (fresh pairing → simulated restart → re-dial connects → a fake
+  Amber answering as a different key is refused) and 3 new vault unit
+  tests (plaintext roundtrip, encrypted fail-closed, legacy-vault
+  parsing). `cargo clippy --all-targets` 0 warnings, `cargo fmt
+  --check` clean, `cargo build --release` green. Frontend untouched.
+
+## Commits added (newest first)
+- `0982dad` feat(nip46): restore saved signer sessions on startup/unlock — no fresh scan
+
+## How to resume / reproduce
+- Restart Electron (new release binary). With the vault already
+  unlocked/unencrypted, the backend logs `[NIP46] restoring session:
+  peer=... as npub1...` then `identity check on restored session:
+  PASS` and the profile goes Connected without showing Amber a new
+  scan. CLI trace: `~/Tools/keynctr-debug/` logs.
+- Then do the still-pending live proof: Publish name / publish a note
+  and approve in Amber within 2 minutes (120s leash from `f53bc56`).
+
+## Outstanding / next steps
+1. **Live sign/publish through real Amber** (covers both the 120s
+   leash and the restored session in one shot).
+2. Prune the 11 stale profileless `nip46_connections` rows (cosmetic;
+   restore already skips them).
+3. Remote picture/nip05 edits, KDF upgrade (Step 5), rename pass
+   (Step 7) — unchanged.
+
+---
+
 # Checkpoint — sign_event given the human-approval timeout leash (2026-09-23 evening)
 
 ## Where things are