Finlynq vs Maybe (and the Sure fork)

When to choose Maybe / Sure

Maybe / Sure is the right call if any of these matter more than ownership:

You want the inherited brand recognition and net-worth-first dashboard from the well-known open-sourced-after-shutdown project.
You want Plaid bank aggregation already in the codebase — Maybe was Plaid-native (plus SimpleFIN). Note: Plaid is dormant after archival, so live sync requires your own setup.
You prefer a conventional Ruby on Rails + Hotwire monolith — easy for senior Ruby contributors to work in.
You want full PFM scope (budgets, transactions, holdings, multi-currency) in one open-source app and do not need first-party AI/MCP.
You specifically want the actively-maintained Sure fork's community momentum.

When to choose Finlynq

Finlynq is the right call if any of these matter:

You want a single-team-led project with a clear roadmap. The original Maybe is archived; Sure is volunteer-run with thinner roadmap clarity.
You want a first-party MCP server. Neither Maybe nor Sure ships one — Maybe's old in-app AI was in-process OpenAI function-calling (not MCP, not exposed to external clients). Finlynq ships first-party MCP (HTTP + stdio) with read and write tools.
You want per-user envelope encryption where the operator cannot read your data. Maybe/Sure has no app-layer column encryption.
You want a modern React / TypeScript stack the AI-tooling ecosystem speaks natively, rather than a Rails / Hotwire monolith.
You want an official managed cloud. Sure is self-host only; Finlynq offers finlynq.com/cloud.
You want to avoid Plaid dependency risk — the same aggregator cost that helped sink the original Maybe company. Finlynq's import + connector framework is Plaid-independent.

Side-by-side

	Finlynq	Maybe / Sure
License	AGPL v3	AGPL v3 (same as Finlynq)
Hosting	Self-host (Docker + PostgreSQL) or managed cloud	Self-host via Docker (Sure fork); original Maybe cloud is gone
First-party MCP	Yes — 91 HTTP / 87 stdio tools	No — Maybe's in-app AI was in-process OpenAI function-calling; no documented Sure MCP
MCP auth	OAuth 2.1 + DCR, Bearer API key, or stdio	N/A (no MCP)
REST / HTTP API	Yes — full surface mirrored from MCP	Partial / internal; no mature documented public REST surface
Bank sync	File / email import + connector framework. No first-party Plaid today.	Plaid + SimpleFIN in code; Plaid dormant post-archival (self-config required)
Encryption at rest	Per-user envelope encryption (AES-256-GCM, scrypt-derived KEK). Operator cannot decrypt.	None at the app / column level
Multi-currency	Native, per-currency cost basis, FX locked at trade date	Yes (native), via a flat exchange-rate table
Investment / portfolio	Lot-tracked cost basis, dividends, FX-aware aggregation; RRSP/TFSA/RESP	Yes — securities, holdings, crypto, performance (a Maybe strength)
Native mobile app	React Native (Expo) app — functional, not at parity with consumer apps	No (Flutter companion code existed but never fully shipped)
Multi-user / household	No (single-user)	Yes (per-instance)
Pricing	Donation-based; same features on self-host and managed cloud	Free (self-host); original paid tier is defunct
Funding / revenue model	Bootstrapped, donations	Original raised ~$1.45M, shut down 2023, open-sourced, then archived July 2025 on a B2B pivot; Sure is volunteer / donation

Migrating from Maybe / Sure

Export from Maybe/Sure — use the CSV export of transactions and holdings, or your database export if self-hosting.
Import into Finlynq at /import/reconcile — review and edit each row; multi-currency, transfer pairs, and dedup are handled in staging; record holdings via the portfolio flow.
Connect Claude (or any MCP client) at /mcp — paste the URL into Claude → Customize → Connectors; OAuth handles auth.

FAQ

Is Maybe still being developed?▾

The original maybe-finance/maybe repository was archived in 2025. The community fork Sure (we-promise/sure) is the de facto continuation and is still active, but it is volunteer-run with a less defined roadmap.

Does Maybe/Sure have an MCP server?▾

No. The original Maybe had an in-app AI chat using OpenAI function-calling internally — not MCP, and not exposed to external AI clients. No documented Sure MCP exists. Finlynq ships first-party MCP.

Maybe has Plaid bank sync — does Finlynq?▾

Maybe was Plaid-native, but Plaid is dormant after archival and needs self-configuration. Finlynq has no first-party Plaid and uses import plus a connector framework instead (SnapTrade on the roadmap).

Both are AGPL v3 and full PFMs — what's the real difference?▾

Project health (single-team-led with a roadmap vs a volunteer fork), stack (React/TypeScript vs Rails/Hotwire), per-user envelope encryption, and a first-party MCP server with write access — none of which Maybe/Sure offers.

Why did Maybe shut down?▾

Per its founders, a combination of bad timing, a long build before validating fit, few paying customers at launch, and painful, expensive Plaid aggregation. Finlynq's donation-funded, Plaid-independent model is a deliberate response.

Sources

maybe-finance/maybe (archived)— archived 2025, AGPL v3
we-promise/sure (community fork)— active community fork, fetched 2026-05-29
Failory — why Maybe failed— post-mortem: funding, Plaid cost
Finlynq on GitHub— AGPL v3 source

Try Finlynq

Free, open source, AGPL v3. Run it on our managed cloud or self-host with one Docker Compose file. Same features either way.

Try the managed cloud Self-host with Docker Source on GitHub