We don't grow at any cost.
Mesh is run by three engineers and funded by customers. We'll stay small. You'll recognise the same names on support tickets in five years.
Mesh is a workflow engine you write in plain Python. It runs reliably, logs clearly, and gets out of the way. There's no dashboard to learn, no retainer to pay, and no sales call required.
// pip install mesh-workflow — and that's the whole setup. from mesh import workflow @workflow(schedule="0 9 * * *") async def daily_reports(ctx): return await ctx.fetch("rows").then(send_email)
We've worked at the other kind of startup. This one is the opposite.
Mesh is run by three engineers and funded by customers. We'll stay small. You'll recognise the same names on support tickets in five years.
If a feature can't be explained in a 500-word doc page, it doesn't ship. If it needs a 30-minute onboarding call, we've already failed.
Self-hosted is a first-class option, not a tier you negotiate for. The same binary, the same docs, wherever you run it.
Every release note lists what broke, what we undid, and what we can't yet explain. No triumphalism.
There is no enterprise-tier feature that's technically ready but artificially withheld. Everyone gets the same product.
We deliberately don't have a long feature page. This is all of it.
Workflows that survive deploys, restarts and network flakes. Guaranteed to finish.
Every step logged, every retry attributed. OpenTelemetry export built in.
Python, TypeScript, Go. You can mix them in one workflow if you need to.
Pin workflows to a region, or distribute them. Same API.
Helm chart, Terraform module, the same binary we run in managed.
Replay any past run. Backfill historical state without rewriting code.
The full price list fits on a page.
Three engineers read every enquiry. No chatbots, no lead qualifiers.
We're a small team across London, Porto and Lisbon. Real reply within a working day — usually much sooner.