Yes, it is suitable for starting development: the README specifies the Ruby, PostgreSQL, Redis, and startup requirements, while clearly separating development from production self-hosting.

• Ruby is the main language. The README says to use the version from .ruby-version, PostgreSQL greater than 9.3, and Redis greater than 5.4, with the latest stable versions recommended.
• The setup flow copies .env.local.example, runs bin/setup and bin/dev, and serves the application at http://localhost:3000.
• The README explicitly labels this section as being for developers contributing to the application; self-hosting users are directed to the Docker guide instead.
• rake demo_data:default can optionally load demo data for basic validation.

This is enough for the Ruby application path. The excerpt does not fully describe separate workflows for the Dart mobile code, JavaScript assets, or Rust code.

Yes, but only if the fork follows the README’s attribution and trademark requirements. The permission to fork does not grant permission to keep using the Maybe name or logo.

• The README states that users may fork the repository under AGPLv3 and must include the original AGPLv3 license.
• A published fork must clearly state that it is based on Maybe Finance and is not affiliated with or endorsed by Maybe Finance Inc.
• “Maybe” is a trademark of Maybe Finance Inc.; forked repositories may not use that name or logo. The README says that “Sure” is not that trademark.
• The License and Trademarks section confirms that both Maybe and Sure are distributed under AGPLv3, so modification and redistribution carry continuing license obligations.

The technical fork is permitted. Compliance depends on preserving license notices, making accurate attribution, choosing a non-infringing brand, and reviewing AGPL obligations for a network service or closed extension.

No, not for the stated constraint: the No AI deployment is the only documented option that avoids an AI token, while the OpenAI and External AI paths add external dependencies and unresolved data-governance questions.

• “Sure - No AI” intentionally leaves AI tokens blank and deploys Sure web, a Sidekiq worker, Render Postgres, and Render Key Value.
• “Sure - Simple AI” uses OpenAI-backed AI, asks Render for an OpenAI token, and provides pgvector-ready Postgres.
• “Sure - External AI” adds external assistant settings and AlphaClaw, which manages the OpenClaw gateway on Render.
• The README lists these deployment variants but does not explain which financial fields are transmitted, whether data is redacted, retention periods, or assistant permission boundaries.

Choose No AI when external transmission and API-key costs are unacceptable. If LLM access is mandatory, the available documentation is insufficient to establish that the privacy requirements are met.

It depends: Sure explicitly acknowledges performance problems in data-heavy applications, so the presence of PostgreSQL, Redis, and Sidekiq alone does not demonstrate stable behavior at large scale.

• The README links to a public performance dashboard showing problematic requests and stack traces from the demo site, and invites contributions to improve performance.
• Local development requires PostgreSQL and Redis, while the Render deployment options include a Sidekiq worker and Render Key Value, making background jobs and caching part of the runtime design.
• The project is still at v2026.11.10-alpha with only five releases, so performance and compatibility work may still be evolving.

If your dataset resembles the README’s data-heavy cases, the available material cannot justify calling Sure a reliably scalable replacement. There is no documented capacity limit, benchmark, or slow-request threshold.

Yes, for access-path coverage: the README explicitly lists all five requested interfaces, although it does not promise identical functionality or stability across them.

• The Hosting Sure section names browser access, a macOS desktop app, a mobile app, API clients, and LLM agents.
• It links to a separate “Sure Clients” document, indicating that these clients are part of the intended project surface rather than integrations users must build entirely themselves.
• The repository contains Ruby, HTML, Dart, JavaScript, and Rust. Dart is consistent with a mobile client, but language presence alone does not prove complete feature parity.
• The latest release is alpha and the project has five releases, leaving client compatibility and release synchronization uncertain.

Sure clearly matches a requirement for multiple access paths. The README is not sufficient evidence that every client is production-ready or feature-equivalent.

It depends: Sure fits a self-hosting and data-control requirement, but its maintenance status and alpha release introduce long-term risk.

• The README explicitly states that Sure can be “self hosted with Docker” and accessed through a browser, macOS desktop app, mobile app, API clients, and LLM agents.
• It is a community fork of the abandoned Maybe Finance project. The latest release is v2026.11.10-alpha, and the project has only five releases, so it should not be treated like a mature commercial service.
• The README also discloses performance issues in data-heavy applications, which may matter as financial records grow.

Sure is practical if you can own database backups, upgrades, and recovery. It is not a low-maintenance consumer service, however, and the README does not establish a long-term support or migration guarantee.