Skip to content
repoless
Now live · crash-aware deploys

Ship apps. Watch them.

Without the platform tax.

A self-hosted PaaS and observability platform — built together. Deploy to your own servers, with crash-aware revisions and log search wired in from the first push.

~/acme · zsh

$

Two products. One platform.

Deploy without ops. Watch without surprises.

Most teams stitch deploys to a separate error tracker and hope they speak the same language. Repoless ships both halves as one product.

deploywatch

Self-hosted PaaS for your server fleet

Multi-host placement with Wireguard mesh, blue-green deploys with Cloudflare-edge holds, and managed databases — driven by a Go agent installed in one curl on any Linux host.

  • Compose, Dockerfile, static, or Nixpacks
  • PR previews + multi-host placement
  • Managed Postgres, Redis, Patroni
  • Cloudflare DNS + zero-downtime swaps

logwatch

Multi-tenant observability with deploy DNA

Real-time ingest, full-text search, auto-grouped issues, and alert rules — every event tagged with the deploy revision that produced it.

  • Full-text log search across services
  • Auto-grouped issues with revision tags
  • Saved queries + collections
  • Slack, Discord, email, webhook alerts

What's in the box

Everything you'd otherwise glue together yourself.

Each item below is a real shipped feature — not a roadmap entry.

Crash-aware revisions

Every revision card shows error-rate delta vs prior. Bad deploys identify themselves.

Incident runbooks

Triggers on error rate, uptime, or resource. Routes to Slack, Discord, email, or webhook.

Multi-host placement

Pin to a single host or spread replicas across a Wireguard mesh. Per-host capacity.

Managed databases

Postgres, Redis, Patroni clusters with extensions, PgBouncer, and pgBackRest PITR.

Real-time log search

Full-text across all services. Saved queries, collections, project-scoped filters.

Auto-grouped issues

Errors deduped into issues with occurrence counts and the revision they first appeared in.

Cost tracking

Host monthly cost × CPU-share allocation. Per-app dollar badges across every project.

RBAC + 2FA

Project-scoped access, deny-by-default. Scoped API tokens. Built-in WebCrypto TOTP.

Why this matters

Errors aren't orphans when deploys leave fingerprints.

Most teams ship in one tool and look for fires in another. The two never quite agree on what's a deploy. Repoless tags every log line with the revision that produced it — so the connection is structural, not stitched.

  • Revision-tagged logs

    Every log line carries the SHA of the deploy that emitted it.

  • Auto-detected regressions

    Error-rate delta vs the previous revision. Bad deploys identify themselves.

  • One-click rollback from logs

    See an error spike on a revision. Click. Roll back to prior. Same panel.

deploywatch·apierr +340%

c0a3f02 refactor: orders queue

82 errors in last 5mOpen in logwatch
searchrev:c0a3f02status:5xx82
14:02:12500POST /v1/checkout1240ms
14:02:12500POST /v1/orders920ms
14:02:12500POST /v1/checkout1180ms

Grouped issue

ECONNRESET upstream=stripe

first seen on c0a3f02 · 82 occurrences

From zero to streaming

Three steps. Around four minutes.

Bring your own GitHub repo, your own servers, your own domain — we handle the connective tissue.

01

Connect a host

Generate an install one-liner. The Go agent hardens the box and joins the Wireguard mesh.

02

Deploy an app

Point at a GitHub repo. Push to main. Health-gated blue-green swap, no downtime.

03

Watch it run

Logs stream in real time. Issues auto-group. Every event tagged with the deploy revision.

Under the hood

Boring infrastructure. Predictable behavior.

No proprietary runtime. No new abstractions to learn. Just the components your team already understands.

Compute

Any Linux host

Edge & DNS

Cloudflare Workers

Mesh

Wireguard private net

Ingress

Traefik + Let’s Encrypt

Common questions

The honest answers.

Coolify is excellent at deploys; it doesn't do observability. Repoless ships the deploy half and the watch half as one product, sharing auth, RBAC, and notifications. We also support multi-host placement with a Wireguard mesh, blue-green swaps with Cloudflare-edge holds, and adoption of existing Patroni clusters. We built a non-destructive Coolify importer if you want to try us alongside.

Those work if you already have a deploy story. Repoless is the deploy story and the observability layer — every log line tagged with the revision that emitted it. Self-hosted, no per-event pricing.

The control plane (deploy.repoless.com, logs.repoless.com) runs on Cloudflare Workers + D1, managed by us. The data plane — your apps, databases, and logs — runs on your own servers. You bring the infra; we drive it.

Any Linux server with SSH access — VPS, dedicated, or bare metal. The Go agent installs in one curl on Ubuntu/Debian and joins the Wireguard mesh. Storage: R2, S3, SFTP, or any S3-compatible object store. DNS: Cloudflare.

Per-team Cloudflare and provider tokens are encrypted at rest with WebCrypto AES. API tokens are dwa_* bearer tokens with scopes. TOTP 2FA is a custom ~80-LOC implementation — no external auth dep.

Yes. The Coolify import path is non-destructive — projects, apps, and managed databases over HTTP. You can also adopt existing external Postgres or Patroni clusters by URL.

We're finalizing it. Free tier for personal projects, paid tier for teams. Get on the list and we'll send early-access pricing before public launch.

Stop renting a platform.

Start owning one.

Three minutes from a fresh Linux host to your first deploy streaming logs back. No credit card, no Kubernetes.