Sure (Free, Open-Source Personal Finance App): What It Is, and What We Tested

Sure is a free, open-source personal finance app — the community fork of Maybe Finance, the app a funded startup spent nearly $1 million building before archiving it. You self-host it with Docker, your data stays on your server, and there is no subscription. We tested it on November 18, 2026, on its public demo (sample data, no real money involved): we created a manual account and a $25 test transaction, checked budgets, rules, CSV import templates and all 24 bank providers, and deleted everything we created afterwards. Every number below is what the app actually showed.

1. What Sure is — and why it is hot right now

Sure (we-promise/sure on GitHub, website sure.am) is an all-in-one personal finance app: track bank accounts, investments, crypto, property and loans in one place, see your net worth, record transactions, set budgets by category, and import history from other apps. It is the community fork of Maybe Finance: the Maybe team built the app through 2021–2022, stopped development in mid-2023, and archived the repository at its final v0.6.0 release in July 2025. Volunteers picked the code up the same month as Sure and have shipped 21 stable releases since; the latest stable release is v0.7.6-hotfix.1 (November 12, 2026), and the public demo runs v2026.11.17-alpha.

The numbers behind the surge: 16,425 GitHub stars as of November 18, 2026 — up from about 13,800 on November 13 and 15,632 on November 15, roughly 2,600 stars in 5 days — and a place at #4 on GitHub's weekly trending list (it also hit #1 on the Ruby daily list on November 10). The licence is AGPL-3.0: free to use and self-host, with the source open for anyone to read. The demand behind it is easy to see in search data: "personal finance app" draws about 12,100 searches a month, "cancel subscription" about 40,500 a month and up 111% year over year, and "no subscription" is a breakout query — people want one place for budgets, accounts and goals, and they do not want another monthly fee. YNAB charges $109/year and Monarch Money $99.99/year; Sure's self-hosted version charges $0. That is the whole pitch, and it is why we tested it for the small-business owners, freelancers and creators who read this site — the same people our Debt Freedom Planner spreadsheet is built for. A spreadsheet plans your payoff; Sure tracks the accounts day to day.

2. Try it — the demo first, then self-hosting

Path A — the public demo (tested by us, $0, no signup). Open app.sure.am. The login page says, verbatim: "This is a demonstration environment. Login credentials have been pre-filled for your convenience. Please do not enter real or sensitive information." The email is pre-filled (user@example.com), the password is masked, and one click on "Log in" takes you into a populated demo with 28 sample accounts and 8,063 sample transactions. No registration, no email, no API key. That is exactly what we did in section 3 — and what we recommend you do first, before you install anything.

Path B — self-host with Docker (official documented steps; see the honesty box below). The official guide (docs/hosting/docker.md) and the example compose file we fetched from the repository on November 18, 2026 describe the same three steps: download the example file, create a .env with two secrets, and start the stack.

mkdir -p ~/docker-apps/sure && cd ~/docker-apps/sure
curl -o compose.yml https://raw.githubusercontent.com/we-promise/sure/main/compose.example.yml
# in .env: SECRET_KEY_BASE from "openssl rand -hex 64", plus POSTGRES_PASSWORD
docker compose up -d
# then open http://localhost:3000

Requirements (from the official docs): a server or computer with Docker and the Compose plugin; the stack runs the Sure web app plus a worker, PostgreSQL 16 and Redis; the docs name about 512 MB of RAM as the lower limit (1 GB with the AI assistant on). On a fresh install you land directly on a "Create account" page — there is no separate sign-up step — and the first account becomes the super admin, so close signups afterwards (Settings → Self-Hosting → Onboarding, or set SURE_ONBOARDING=closed). The compose file also includes a daily backup service that keeps seven days.

Honesty box — what we did and did not run. We verified the demo end to end (section 3), but we could not run the Docker install to completion on our test machine, and we will not pretend otherwise. Our sandbox blocks the three things Docker needs: iptables (daemon refused with a permission error), overlayfs mounts (layer extraction failed), and starting any container at all — even with --privileged — because the sandbox forbids creating network namespaces. Docker Hub also returned EOF repeatedly from this machine (the mirror at mirror.gcr.io worked, and the Sure image itself pulled fine from ghcr.io). The steps above are therefore the official documented steps, cross-checked against three independent guides that each ran them for real (deployn.de installed Sure 0.7.6 with Docker Compose on a Debian server; lockergnome and Portainer templates describe the identical two-secrets-and-one-command flow) — not steps we executed ourselves. If you run them on a normal server or VPS, they are the path; if they fail for you, the demo in Path A still shows you the whole app first.

3. Our hands-on test: the public demo, step by step

Target: prove — on the real app, with sample data we could check in advance — that the core loop works without connecting a bank: create an account by hand, record a transaction by hand, and see budgets move. Everything below was done on the demo at app.sure.am on November 18, 2026; the test account and test transaction were deleted afterwards, and the demo was left as we found it.

Tool / version testedSure v2026.11.17-alpha
GitHub stars16,425 (+~2,600 / 5 days)
LicenceAGPL-3.0
Cost$0 (demo + self-host)
Demo data28 accounts, 8,063 txns
Date tested2026-11-18

Test 1 — Dashboard: $537,283.14 net worth

After the pre-filled login, the dashboard showed Net Worth $537,283.14 (1-day view: +$337,164.92, 168.5%; 5-year view: +$944,249.14, 232.3%), Assets $707,781 and Liabilities $345,612: Cash $192,374.40, Investments $165,806.21, Crypto $100, Properties $325,000, Vehicles $22,500, Other $2,000; Credit Cards $7,404, Loans $337,409, Other $800. Last 30 days: Income $30,262.15, Spending $13,199.10 (against a $3,000 budget), Savings Rate 56.4%. The sample data spans currencies — USD, EUR €29,900 (Deutsche Bank), HUF 38,745 Ft (Erste Bank), GBP £14,796.80 (Vanguard UK ISA) — and familiar merchants: Whole Foods, Delta Air Lines, Starbucks, Airbnb, Amazon, Walmart, Shell, Uber, Netflix. Honestly recorded: the demo's own numbers do not add up — $707,781 − $345,612 = $362,169, not the $537,283.14 net worth shown. That is a demo-data inconsistency, not a calculation we can blame on the app's maths, but it is worth knowing before you trust any sample screen at face value.

Test 2 — A manual account: created, then deleted

There is no standalone "Accounts" navigation; the list lives in the homepage sidebar and at /accounts. Asset types are Cash, Investment, Crypto, Property, Vehicle and Other Asset; liabilities are Credit Card, Loan and Other Liability; Cash splits into Checking, Savings, Money Market and CD. We created a test account via "New cash → Enter account balance": name TEST-DELETE-ME, type Savings, balance $1,234, opening date defaulted to 2024-11-18. It worked: the app confirmed "Depository account created", Cash rose from $192,374.40 to $193,608.40, and the account appeared in the sidebar. Deleting it (More → "Eliminar cuenta") scheduled the deletion rather than removing it instantly — the notice reads "Cuenta Depository programada para eliminación" — and Cash returned to $192,374.40.

Test 3 — A manual transaction: $25, then deleted

From the Transactions page we recorded an Expense: "TEST-DELETE-ME expense", account TEST-DELETE-ME, $25.00, category Groceries, date 2026-11-18. After the save, the totals moved exactly as they should: transactions 8,063 → 8,064, Expenses $1,017,068.96 → $1,017,093.96, and the test account balance fell to $1,209.00. A search confirmed the entry (Nov 18, 2026, Groceries, −$25.00, "Protected from sync"). Deleting it restored the balance to $1,234.00. One real bug on the way: the "New → New transaction" button on the account page did nothing on two clicks — we had to create the transaction from the Transactions page instead.

Test 4 — Budgets: monthly limits, not zero-based

The sidebar "Plan" link did nothing when clicked; going directly to /budgets redirected to /budgets/oct-2026 (period Oct 24 – Nov 23, 2026). The demo's budget was $500 total, with $20,197.95 spent — 14 categories over, 3 on track. Food & Dining spent $1,521 against a $250 limit; Housing $7,866.62 against a $0 limit; Uncategorized $5,000 against $250. The edit page showed only two non-zero limits in the whole budget (Food & Dining $250, Uncategorized $250), each category with a Rollover switch and an average-spend figure — some plainly wrong in the sample data (Housing " $70,000/mo"). The important finding is structural: Sure's budgets are monthly spending limits per category, with optional rollover — not YNAB-style zero-based envelope budgeting, where every dollar is assigned before it is spent (see FAQ).

Test 5 — Filters and rules: both work

The Transactions page lists every filter we would want — Account, Date, Type, Status, Amount, Category, Tag, Merchant, AI, Attachment — and filtering by Category = Groceries plus a search for our test entry returned exactly 1 result ($25.00). The rules page (/rules) holds 5 rules, all enabled: a merchant matcher, a category classifier (both marked "no AI"), and three mapping rules (a café name → Coffee & Takeout, WholeFoods → Groceries, Delta Airlines → Travel), applied to past and future transactions, top to bottom. The last run was logged as Nov 18, 2026, 02:51–02:52 PM, manual, successful.

Test 6 — CSV import: templates confirmed, a real file not run

The import page (/imports) offers templates for Sure (.ndjson full export), Mint, Actual Budget, Quicken (QIF) and YNAB, plus an AI document import (PDF and other files) and raw CSV imports for transactions, investments, accounts, categories, merchants and rules. The generic transaction flow has five steps — Upload → Configure → Clean → Map → Confirm — with comma/semicolon separators, an account picker and a sample CSV download. We walked the flow but uploaded no file, and we cannot claim a real import succeeds: the demo's own 11 existing import records were all stuck at "En progreso" (from 09:31 to 17:24 that day), with none completed. If migrating is your reason to try Sure, test the import with a small file first.

Test 7 — Bank connections: 24 providers, no Plaid in the list

The providers page (/settings/providers) lists 24 available providers and 1 connection (Apple Wallet Beta, sync active, linked to an Apple Card $829.76, Apple Cash $1,913.01 and "Nancy's" $17.50). The 24 are Akahu, Binance, Brex, Coinbase, CoinSpot, CoinStats, Enable Banking, Fio banka, Indexa Capital, Interactive Brokers, Kraken, Lunch Flow, Mercury, Monobank, On-chain wallets, Questrade, Redbark, SimpleFIN, SnapTrade, Sophtron, Trade Republic, Trading 212, Up and Wise — most marked Beta or Alfa. Plaid is not in the list (the in-app guides still mention Plaid for US and EU, so the list and the docs disagree — check before you count on it). Opening SimpleFIN showed the pattern: read-only, credentials encrypted, and it asks you to paste a one-time setup token from your own SimpleFIN Bridge. Most providers want your own read-only API key or token; Sure itself charges nothing for syncing, but whether a provider charges you is the provider's business, not Sure's. We entered no keys anywhere.

Test 8 — AI assistant and settings: what we found

The settings area is split across sub-pages (/settings itself returns a 404 — a real bug). The AI prompts page shows an enabled assistant with four editable prompt groups (main system, transaction categorizer, merchant detector, insight dig-in). Usage for Oct 19 – Nov 18, 2026: 30 requests, 513,724 tokens, total cost $0.00, model @cf/moonshotai/kimi-k2.5. The transaction classifier is a Bayes classifier with editable priors and thresholds — not an OpenAI setup — and the "API keys" page is for Sure's own programmatic keys, not for an OpenAI key. In the pages we viewed, no field asked for an OpenAI-compatible key: on the demo, the AI model is configured server-side. Note the difference for self-hosters: the compose file includes an OPENAI_ACCESS_TOKEN variable, and its own comment warns that enabling OpenAI "will incur costs" — on your own server, the AI assistant is bring-your-own-key.

What we checkedResult
Demo login (pre-filled, sample data)Passed. One click, no signup, no real information entered. Warning text quoted in section 2.
Dashboard / net worthPassed, with a data caveat. Net worth $537,283.14 shown; assets minus liabilities = $362,169, so the demo's own numbers disagree.
Manual account ($1,234 Savings)Passed. Cash $192,374.40 → $193,608.40; deleted afterwards, Cash restored.
Manual transaction ($25 Groceries)Passed. Totals 8,063 → 8,064, expenses +$25.00, balance $1,209.00; deleted afterwards. Account-page "New transaction" button dead — use the Transactions page.
Budgets (monthly limits)Passed. Limits editable, rollover switches present; $500 budget vs $20,197.95 spent in demo data; not zero-based budgeting.
Filters + rulesPassed. Category filter exact; 5 rules enabled, last run logged successful.
CSV import templatesPartly verified. Templates confirmed (Mint / YNAB / Actual Budget / QIF / .ndjson / AI doc); no file actually imported; demo's 11 imports all stuck "En progreso".
Bank providers (24)Passed (viewed only). No Plaid in the list; most need your own read-only key/token; no keys entered.
AI assistant / settingsPassed (viewed only). 30 requests / 513,724 tokens / $0.00 on the demo; Bayes classifier; no OpenAI key field seen on the demo.
Cost, account, API keyPassed. $0 on the demo; self-host free under AGPL-3.0; bank sync and AI keys are bring-your-own.

Our verdict: the app itself is real and the manual-tracking loop works exactly as advertised — no bank connection needed, no subscription, no account to create just to look. The demo server is the weak part (see section 4), and two things we could not verify — a Docker install on our machine and a real CSV import finishing — are exactly the two things to test yourself before you move your real finances in.

4. What Sure can't do — the limits, stated plainly

5. Safety — read this before you run it

6. FAQ

Is Sure really free? What's the catch versus YNAB ($109/year)?

Yes, for the self-hosted version: AGPL-3.0, no subscription, no per-account fee — you pay only for the server it runs on (or $0 if you already have one). The catch is not a hidden fee, it is work: you install and maintain it, you bring your own keys for bank sync and for the AI assistant, and you get a volunteer-maintained app with rough edges instead of a polished commercial one. YNAB's $109/year buys you zero-based budgeting, a polished mobile app and support; Sure buys you net-worth tracking, investments and freedom from the subscription, with a different budgeting method (see below). One unresolved detail: some write-ups describe a hosted version at app.sure.am as free for 45 days and then $2.50/month or $25/year, while others say no proper hosted version is offered yet. What we actually saw was a demo in "Demo Mode" with pre-filled credentials and no signup flow, so we cannot confirm any hosted price — treat self-hosting as the free route and verify any hosted offer yourself before paying.

What happened to Maybe Finance? Is it dead?

Effectively, yes. The Maybe team built the app in 2021–2022, stopped development in mid-2023, and archived the repository at the final v0.6.0 release in July 2025 — it is read-only now, nobody fixes its bugs or security issues, and its stock prices and exchange rates came from a single outside provider. Sure is the community fork that picked the code up in July 2025: its first release was a copy of Maybe v0.6.0, and it has shipped 21 stable releases since, the latest v0.7.6-hotfix.1 on November 12, 2026. If you want Maybe back, Sure is where its code went — but it is a community project, not the original company.

What do I need to self-host it?

Docker and the Compose plugin, plus about 512 MB of RAM (1 GB with the AI assistant on). The official flow is: download compose.example.yml, create a .env with SECRET_KEY_BASE (from openssl rand -hex 64) and POSTGRES_PASSWORD, run docker compose up -d, and open http://localhost:3000. The stack includes PostgreSQL 16 and Redis. Two things to do immediately after: create your account (the first one is the super admin) and close signups. One honest caveat: we could not run this install on our test machine — our sandbox blocks Docker containers entirely (see the honesty box in section 2). The steps are the official documented ones, corroborated by independent guides that ran them for real, but verify them on your own server.

Can it sync my bank? Does it cost extra? What about Canadian banks?

It can sync through 24 providers listed in the app — including Wise, Questrade (Beta), SimpleFIN, Enable Banking, Lunch Flow, Mercury, Trade Republic and several crypto exchanges — but each provider needs your own read-only API key or token, and whether a provider charges you is up to that provider; Sure itself charges nothing for syncing. Two warnings from our test: Plaid is not in the demo's provider list even though the in-app guides mention it, so do not assume Plaid works; and we did not test any live bank connection, Canadian or otherwise, so we cannot confirm a specific Canadian bank syncs. Manual entry and CSV import need no provider at all, and that is the path we verified.

Where does my financial data live, and is it safe?

On a self-hosted install, your data lives in your own PostgreSQL database, on your own server — no third party holds your accounts. That is the privacy case for self-hosting, and the reason the app asks you to protect your SECRET_KEY_BASE like a password-manager master key. On the public demo, the opposite is true: it is a shared environment with sample data, and its own warning says not to enter real or sensitive information. The rule is simple: demo for trying, self-host for real data — and always back up before upgrades.

How is Sure's budgeting different from YNAB's?

YNAB is zero-based envelope budgeting: every dollar gets a job before you spend it, and the app is built around that discipline. Sure is broader and looser: net worth comes first, and budgets are monthly spending limits per category, with an optional rollover of unspent money. In our test the demo budget had only two non-zero limits in the whole month, and most categories sat at $0 while spending ran into the thousands — the app did not force us to assign anything. If you want YNAB's method without the fee, Actual Budget (envelope budgeting, open source) is the closer match; if you want net worth, investments and budgets in one place and can live with limits instead of envelopes, Sure is the fit.

Can I migrate from YNAB, Mint or Actual Budget?

The templates exist: the import page offers Mint, YNAB, Actual Budget, Quicken (QIF) and Sure's own .ndjson export, plus an AI document import and raw CSV imports for transactions, investments, accounts, categories, merchants and rules. But we did not verify a real import end to end — we walked the five-step flow without uploading a file, and all 11 of the demo's imports were stuck at "En progreso". So the honest answer is: the door is open, but test it with a small export first (50 rows, one account) and confirm the balances match before you migrate everything.

Want your money tracked the same way?

A tracker shows you where the money is — a plan decides where it goes next. Our Debt Freedom Planner is the companion spreadsheet: pick Snowball or Avalanche, watch the debt-free date move, and see exactly how much interest extra payments save (CA$24.99). And if the business side needs the same treatment, see our free tools — including a quote generator and a printing cost calculator — and our AI Business Pack, a bilingual website with 24/7 AI customer service from C$699. If you want help setting any of this up, talk to us.

Want it printed?

We can put your design on T-shirts, hoodies, mugs, posters and more — printed to order in Quebec, no minimums, with a free preview before you pay.

Contact us for a free quote