返回资讯中心
外部精选
软件工程
#开发者实时流

Show HN: SiteSolide – Open-source cloud for small software

Internal tools, dashboards and prototypes, written by your people and their agents, go live on HTTPS behind

Hacker News 最新ClemDev20002026/10/19 · 27 天前9 分钟阅读

以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。

Internal tools, dashboards and prototypes, written by your people and their agents, go live on HTTPS behind your company's sign-in with one command.

$ curl -fsSL https://github.com/cthiriet/sitesolide/releases/latest/download/install.sh | sh

CopyCopied

Get started Star on GitHub

licenseMIT

releasev0.6.2

CLIsingle binary

svg]:-mt-[8%] etroit:[&>svg]:-mb-[2%]" aria-hidden="true" data-illustration>

ONE SERVER YOU OWN debian 13

sales.acme.com crm.acme.com pricing.acme.com agent worker digest next run 08:00 company sign-in

apps at their own address

a worker

a scheduled job

your sign-in

Everything a small app needs

Deploying, signing in, sharing, background work, backups and monitoring come built in, on a machine you own.

Run sitesolide deploy in a folder: a static site, or an app in Bun, Node, Python, Go or any binary. It builds, installs, restarts, then checks the app answers on HTTPS.

One machine you rent, at Hetzner or any Debian 13 server. Caddy serves every app on HTTPS, at its own subdomain or on your domain, and systemd runs them. Apps, data and secrets stay on it: no containers, no per‑seat bill.

Your company's sign-in

Google Workspace, Entra, Okta or any OpenID Connect provider in front of private apps, which write no sign-in code and learn who came in.

Open an app to a colleague by their work email, or to everyone at your domain, from the CLI or the dashboard. It holds from their next request.

Revoke

Team tokens, not root SSH

Colleagues and agents deploy over HTTPS with a personal token, scoped and revoked in one click. Only the owner holds SSH.

3041

webserver"port": 3041

agentworker"internal": true

digestscheduled"schedule": "Mon..Fri 08:00"

Workers and schedules

An agent that looks for leads around the clock, a report every weekday at eight: declare them in sitesolide.json beside the app. systemd brings a worker back when it falls and runs each schedule on time, and the sites keep the processor first.

Each project's data snapshotted hourly, SQLite copied consistently, optionally to an encrypted bucket. Restore from the dashboard.

Every command speaks --json, one event per line, each error with a hint. sitesolide mcp serves deploy, logs and sharing to Claude Code, Codex or Cursor, and llms.txt indexes the docs.

sitesolide monitor, example.com: 1 downDOWN https://my-app.example.com/ answered 502

sitesolide monitor, example.com: 1 recoveredRECOVERED https://my-app.example.com/ answered 200 (after 8 min)

Every site checked each minute: one message to Slack, Discord, Google Chat or ntfy when it goes down, one when it recovers. A heartbeat notices the machine itself dying.

Different tools for different jobs: a hosted cloud, a dashboard for containers, a container deploy tool.

Sources: each project's own site and docs, checked on 5 October 2026.

Once the CLI is installed: a domain on Cloudflare, and a Hetzner Cloud token or any Debian 13 server you reach as root.

$ sitesolide machine create --provider hetzner --name web

CopyCopied

2

Install sitesolide on it

hardened, DNS, Caddy, dashboard

$ sitesolide setup root@203.0.113.10 --zone example.com --email you@example.com

CopyCopied

3

Deploy a project

live at my-app.example.com

$ cd my-app && sitesolide deploy

CopyCopied

Read the docs What each step checks and changes

正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。

打开原文
去社区讨论 →