Frequently asked.

Both. Project Meru is the initiative and the home for its public writing; Sumeru is the framework it produces. In the code, the CLI, the configuration file and the Go module names you will only ever see sumeru.

You can run it, and the commercial flow works. But it is pre-alpha: there is no tagged release, no upgrade path between versions, no test suite on the business addons, and gaps as basic as tax. Treat it as something to build on and argue with, not something to migrate onto.

Yes. Addons are linked in at build time through a generated file of blank imports, so adding one means running make generate and rebuilding. Installing an already-compiled addon into a database is a runtime action. From the Apps screen or with -i.

A JSON-RPC endpoint at POST /api/rpc, authenticated with an API key you generate from your user record. It exposes search, search_read, read, create, write and unlink.

No. One running server serves one database. What it does support is multiple companies inside that database, with a switcher and per-user company access. Serving separate customers means separate deployments.

PostgreSQL, and nothing else. Scheduled jobs and the audit trail live there too, so there is no broker or cache to operate alongside it.

A small runtime footprint and a single build artefact remove most of the operational work we set out to avoid. Compile-time types catch a class of bug that matters more in accounting than in most software, and goroutines let one process handle concurrency that would otherwise need a worker pool.

No. Pages are rendered by Go templates and a render package, styled with plain CSS and enhanced by native ES modules. There is no bundler and no build step.

Two. The framework and the standard business addons are open source under Apache 2.0. A separate enterprise edition, under a commercial licence, will add addons and support on top of that same core. Nothing that is open source today will be moved behind the commercial licence.

No. The community edition is a complete system, not a trial: no user cap, no feature flags, no expiry. Neither edition is on sale yet. The project is pre-alpha.

There is a CSV import in the UI, and the tables are plain Postgres, so a bulk import is an ordinary ETL job. Purpose-built migration tooling is not on the near roadmap.

Design feedback helps most right now. If you have run an ERP in production, tell us where it failed.