diff --git a/docs/comparison.md b/docs/comparison.md
new file mode 100644
index 0000000..07161ba
--- /dev/null
+++ b/docs/comparison.md
@@ -0,0 +1,135 @@
+# How BAL compares
+
+!!! danger "Draft — needs review before publishing"
+ This page describes other projects (Nunchuk, Liana) whose features and prices
+ change over time. Everything below was compiled from public sources in **mid-2026**
+ and has **not** been validated by those projects or by the BAL developers. Verify
+ current details on each project's own site before treating any claim as authoritative.
+ Tone rule for this page: **factual, no disparagement** — these are design trade-offs,
+ not mistakes.
+
+Bitcoin After Life is not the only way to plan Bitcoin inheritance. Nunchuk and Liana are
+two well-known, respected projects that solve the same problem differently. This page compares
+them across the dimensions that tend to matter over a multi-year time horizon — decentralization,
+long-term cost, the wallet each one builds on, distribution, and exposure to a single company
+being shut down — and it states plainly where each approach, BAL included, is weaker.
+
+## At a glance
+
+
+
+| | **Bitcoin After Life** | **Liana** (Wizardsardine) | **Nunchuk** (Honey Badger) |
+|---|---|---|---|
+| Mechanism | Pre-signed transaction, absolute `nLockTime`, held by Will-Executors | Script wallet (Miniscript), relative timelocks in the script | Collaborative-custody multisig (2-of-4) with a timelocked inheritance key |
+| Who holds keys | **Only you.** Will-Executors hold no keys | **Only you** (and any recovery keyholders you choose) | You hold 3 keys; **Nunchuk holds 1 platform key** on its servers |
+| Base wallet | Runs inside **Electrum** (in use since 2011) | Own codebase (Miniscript / Bitcoin Core) | Own app |
+| Recurring cost | **None** | **None** for the wallet, but timelocks need periodic coin rotation → network fees over time | **Subscription** ~$120–$480/yr |
+| When you pay | **Only at execution**, from the inheritance transaction, in bitcoin | Network fees on each rotation | Yearly subscription + network fees |
+| Distribution | Desktop plugin (`.zip`) — no app store | Desktop + mobile | Primarily mobile app (App Store / Google Play) + desktop |
+| Single-company shutdown risk | **No company to shut down** | Wallet survives; hosted backend optional | Assisted flow depends on the company; on-chain failsafe as backup |
+| Heir's effort | None — the network broadcasts automatically | Heir signs a recovery transaction after the timelock | Guided beneficiary flow; on-chain failsafe via timelock |
+| Open source | Yes | Yes | Yes (app); platform key is a hosted service |
+
+
+
+## Decentralization
+
+**BAL** is the most decentralized in day-to-day operation: a network of independent
+Will-Executors, with no coordination between them and no company in the middle. In fairness,
+it does depend on Will-Executors continuing to operate, and it uses the [WeList](will-executors/welist.md)
+as a discovery directory — a mild point of centralization you can bypass by adding servers
+manually.
+
+**Liana** is strongly decentralized on the keys — full self-custody, no company key — and its
+recovery path is enforced by the Bitcoin **script itself**, i.e. by consensus rules rather than
+any service. Its convenient hosted backend (Liana Connect) is optional; you can run your own node.
+
+**Nunchuk**'s inheritance plans use *collaborative custody*: the company holds one multisig key
+on its servers. Nunchuk 2.0 (November 2025) added an on-chain autonomous failsafe using timelocks
+and Miniscript, which reduces that dependency — but the platform key remains part of the assisted
+setup.
+
+## No cost that recurs over the years
+
+**BAL** costs nothing until the inheritance is actually executed. There is no subscription and no
+forced periodic action: the only Will-Executor fee is taken from the time-locked transaction
+itself, in bitcoin, and only if a Will-Executor broadcasts it. Keeping the plan current means
+[renewing the will](user-guide/updating.md) yourself — no payment involved.
+
+**Liana**'s wallet is free, but its *relative* timelocks have to be reset periodically by moving
+coins to a fresh address (a Bitcoin consensus constraint, not a wallet flaw). That means paying
+network fees over time and fits cold storage that is never touched poorly.
+
+**Nunchuk**'s inheritance features live behind a yearly subscription (roughly $120–$480 depending
+on the plan). It is a real recurring cost, and the plan lapses if it stops being paid.
+
+## A single, battle-tested wallet: Electrum
+
+**BAL** does not build a new wallet. It runs as a plugin **inside Electrum**, which has been in
+use since 2011 and is among the most heavily reviewed and tested Bitcoin wallets in existence.
+Electrum performs the signing; BAL never signs on your behalf. Electrum is open source and does
+not depend on any single company to keep existing, which matters for a plan meant to work decades
+from now.
+
+The honest counterpoint: BAL *is* a plugin, so it depends on the plugin being maintained and on
+Electrum's plugin interface staying stable. (Note the correct wording: BAL is **listed in
+Electrum's third-party plugin directory** — not "approved by Electrum".)
+
+**Liana** and **Nunchuk** ship their own, newer codebases. Both come from competent teams, but
+neither has Electrum's number of years in the field.
+
+## No dependency on an app store
+
+**BAL** is a desktop plugin installed from a `.zip` file. It never passes through the Apple App
+Store or Google Play, so it cannot be pulled from a store or blocked by a platform's policy — a
+recurring problem for crypto apps in some jurisdictions.
+
+**Nunchuk** is primarily a mobile app distributed through the App Store and Google Play (a desktop
+version also exists), which makes store availability a real dependency. **Liana** is distributed
+from its own site for desktop, with mobile options, so it is less exposed than a mobile-first app.
+
+## Resistance to a company being shut down
+
+This is where an inheritance plan's time horizon really bites: a plan may need to work in ten or
+twenty years, across changes in law.
+
+**BAL** has **no company operating a custody service** — so there is no single entity a government
+can shut down to break the plan. Will-Executors are independent, can be located anywhere, and at
+least one operates Tor-only. If new anti-Bitcoin rules appear in one country (as has been discussed
+in places like Russia), there is no headquarters to close and no license to revoke.
+
+**Liana** is made by a company (Wizardsardine), but the wallet is self-custodial and open source:
+if the company disappeared, existing scripts keep working on-chain because the recovery path is
+enforced by the script, not the company. Only the optional hosted backend would stop, and you can
+switch to your own node.
+
+**Nunchuk** runs a collaborative-custody service with a platform key, which makes the company a
+regulatory target — a hostile jurisdiction could force it to relocate, and closure is at least as
+likely as relocation. To their credit, Nunchuk engineered the on-chain timelock failsafe precisely
+so beneficiaries can still recover **without** the company; but the *assisted* experience does
+depend on the company continuing to operate.
+
+## Where the other approaches are stronger
+
+Fairness requires stating this plainly — BAL is not the best choice on every axis:
+
+- **Liana** gives you a cancellation window by design and needs no external network to broadcast;
+ the recovery path is enforced by the script itself. If you keep coins consolidated, its model is
+ very clean.
+- **Nunchuk** offers a guided, supported experience with a beneficiary flow and a company standing
+ behind it — valuable for users who want assistance rather than a self-run plan, and who are
+ comfortable with one key in collaborative custody.
+- **BAL**'s own trade-offs: it is **desktop-only** (Electrum), it has **no built-in cancellation
+ window** (so it asks the owner for renewal discipline — see [Keeping the will up to date](user-guide/updating.md)),
+ and it relies on Will-Executors continuing to operate (mitigated by redundancy and the
+ [offline backup transaction](user-guide/backup-transaction.md)).
+
+There is no single best choice — only the one that matches how much you want to self-manage,
+whether you accept a third-party key or an app store or a company in the loop, and whether you
+want a plan that costs nothing and depends on nothing but Bitcoin.
+
+---
+
+!!! question "Still have a question?"
+ The [FAQ](faq.md) answers the most common ones — costs, hardware wallets,
+ privacy, and what happens if a Will-Executor disappears.
diff --git a/docs/stylesheets/extra.css b/docs/stylesheets/extra.css
index 9734736..ec1f27c 100644
--- a/docs/stylesheets/extra.css
+++ b/docs/stylesheets/extra.css
@@ -401,3 +401,10 @@ img.diagram + em {
padding-right: calc(100% - 16.1rem);
}
}
+
+/* -------- Compact table (wide comparison tables) -------------------------- */
+/* Default table body is .78rem; 66% ≈ .515rem so a 4-column comparison fits
+ without horizontal scroll. Header cells scale down proportionally too. */
+.md-typeset .compact-table table:not([class]) { font-size: .515rem; }
+.md-typeset .compact-table table:not([class]) th { font-size: .46rem; }
+.md-typeset .compact-table code { font-size: .9em; }
diff --git a/internal/source-notes.md b/internal/source-notes.md
index 3bf0890..7c133e8 100644
--- a/internal/source-notes.md
+++ b/internal/source-notes.md
@@ -52,6 +52,15 @@ Compiled July 2026. Keep this updated whenever the manual is revised.
5. **Fee range currently charged** by Will-Executors on the network (for the recruitment article).
6. **Minimum system requirements** for a Will-Executor node, stated concretely.
7. **BIP 128 critique** — see `design-choices-DRAFT.md`. The single most exposed point.
+9. **`docs/comparison.md` (new, 2026-07-28) — competitor comparison, needs review.** Facts on
+ Nunchuk (Iron Hand ~$120/yr, Honey Badger ~$480/yr, no-KYC/email only, accepts BTC or card,
+ collaborative custody = 1 platform key; Nunchuk 2.0 added on-chain timelock+Miniscript
+ inheritance, Nov 2025) and Liana (free/open-source, Miniscript relative timelocks, self-hosted
+ or Liana Connect hosted backend) were pulled from public web sources mid-2026. **Prices change.**
+ Corrected the user's original premise: Nunchuk is NOT KYC — the honest BAL differentiator is
+ "no recurring subscription + no third-party-held key", not "no credit card". Cross-check with
+ `design-choices-DRAFT.md` editorial rules (no "investment/yield", no disparagement) before
+ publishing. Page carries a draft warning.
8. **`docs/examples.md` (new, 2026-07-28) — needs a full accuracy pass.** Written from
`updating.md`/`delivery-time.md` rather than tested behaviour; user flagged it as containing
"parecchie imprecisioni" without specifying which. Marked with a draft warning on the page
diff --git a/mkdocs.yml b/mkdocs.yml
index 7ed80ca..21d47e9 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -37,7 +37,7 @@ theme:
extra_css:
# The ?v= suffix defeats browser caching of the stylesheet. Bump the number
# whenever you change extra.css and the browser keeps showing the old style.
- - stylesheets/extra.css?v=13
+ - stylesheets/extra.css?v=14
nav:
- Home: index.md
@@ -64,6 +64,7 @@ nav:
- Protocol:
- How it works: protocol/how-it-works.md
- REAL EXAMPLE: examples.md
+ - How BAL compares: comparison.md
- Security & Privacy: security-privacy.md
plugins: