mesh.dev
Status: all systems operational · p99 latency 142ms · 99.97% uptime · 30d

Workflows that survive the real world.

Durable execution, observable by default. Built for engineers who read their logs. Python, TypeScript and Go SDKs. Self-hosted or managed.

~/mesh-workflow · zsh
$ pip install mesh-workflow
Installed mesh-workflow-2.4.0

$ mesh run daily_reports
[09:00:01] INFO  started workflow wf_a1b2c3
[09:00:02] INFO  step fetch_rows → 1,248 rows
[09:00:04] WARN  email_report retry 2/3 (rate_limit)
[09:00:07] OK    wf_a1b2c3 done · 1,248 sent · 6.2s

Built for engineers who ship.

Retries, observability, scheduling, backfills — the patterns every team ends up reimplementing, done once, properly.

cap.01

Durable execution

Workflows survive deploys, restarts and network flakes. Guaranteed to finish.

cap.02

OpenTelemetry

Traces, metrics and structured logs exported to your existing stack.

cap.03

Replay & backfill

Replay any workflow run. Backfill historical state without rewriting code.

cap.04

Multi-language SDK

Python, TypeScript, Go. Mix them in one workflow if you need to.

cap.05

Self-hosted option

Run in your own cluster. Helm chart, Terraform module, zero lock-in.

cap.06

Event-driven

Trigger from webhooks, cron, queues, or another workflow. Your choice.

Per-run. Not per-seat.

Honest pricing that scales with what you actually use.

hobby
$0/mo

Personal projects and early prototypes.

  • 50,000 runs/month
  • Full SDK + dashboard
  • Community support
  • SSO
  • Multi-region
start_free
enterprise
talk

Self-hosted with compliance on.

  • Self-hosted
  • SOC 2 Type II
  • Custom SLAs
  • Dedicated engineer
  • Multi-tenant ready
contact_sales

Drop us a note.

We read every message. Humans only — we don't route enquiries through chatbots.

Office

We're a small team — eight engineers across London, Porto and Lisbon. Real reply in < 24h.

email
[email protected]
sales
[email protected]
github
github.com/mesh-dev
status
status.mesh.dev