Sealos is your ops agent for shipping and running full-stack apps.

Stop stitching Vercel, Supabase, and Railway together by hand.

Deploy with Postgres and Redis
Architecture
Deploy ./bookings with a Postgres and Redis
Thinking
A Next.js app with Prisma and a session store. I’ll build it, add both databases and wire them in.
DetectingNext.js 14 · Prisma
Creatingbookings-pg · PostgreSQL 16
Creatingbookings-redis · Redis 7
InjectingDATABASE_URL, REDIS_URL
Deployingbookings · 41s
Live at bookings.sealos.run. I’ll keep watching its logs and metrics.
Ask Sealos Agent anything…
Default
Services you deploy show up here
DATABASE_URLREDIS_URL
bookings
./bookings
Detecting Next.js
Public URL after deploy
bookings-pg
postgres:16
Creating
bookings-redis
redis:7
Creating
Add service

Turn local code into a live app

Point Sealos at any folder on your machine, from a full codebase to a single HTML deck. It builds the image and puts it on a public URL.

quote-deck Architecture
Publish local ./quote-deck sales deck with live PostgreSQL lead capture.
quote-deckDeployed successfully4/4
Local directory packagedWired leads-pg DATABASE_URL8/8 static assets 200 OKquote-deck.sealos.run
Ask about this plan, or tell me what to change…
https://quote-deck.sealos.run
Interactive quote deck
Growth CloudPopular$49/mo
Dedicated PostgreSQL 16 and custom HTTPS hosting.
Selected · <div.pricing-tier> · 520×96
Agency Multi-tenantIsolated$129/mo
12 isolated client workspaces with daily snapshots.
ROI widgetLiveIncluded
Real-time quote estimator wired to leads-pg.
Add service

Deploy from a GitHub link

Pick a repository and branch. Sealos detects the port, deploys it, and injects the connection string of any database you need.

storefront Building
storefront
acme/storefront · main
Deploying Building image
Public URL after deploy
postgres
PostgreSQL 16
Running
Add service
storefront
acme/storefront · main · a41c9e2
OverviewDeploymentsLogsMetricsVariables
Initial deploy Deploying
a41c9e2 · main · by Jamie Lin · 1m 02s
Clone repository3.2s
Build image58s
Push image
Start container
Health check
Assign public URLport 3000
LATEST LOGSView full logs ›
10:42:11 #5 [deps 2/3] COPY package.json ./ 10:42:12 #6 [deps 3/3] RUN npm ci 10:42:41 added 812 packages in 29s 10:42:42 #9 [build 3/4] RUN npm run build

Everything stays in view

Once it is live, every service sits in one project view with its status, uptime, metrics and logs, so you can open a terminal in any container, roll back a deploy, or catch a memory limit before it is hit.

postgres
PostgreSQL 16
OverviewLogsMetricsSettings
Running1 instance
Uptime
3d 4h
Restarts
0
Last deploy
2h ago
ConnectionEdit in Settings ›
Host
postgres-postgresql.ns-ffmbxs.svc
Port
5432
User
postgres
Password
••••••••••••••••
Connection string
postgresql://postgres:••••@postgres-postgresql.ns-ffmbxs.svc:5432
UsageMetrics ›
CPU24%
Memory62%
Storage12%

Agent keeps it running

After the deploy, your Ops Agent reads logs and metrics, names the cause of a 502 or an out-of-memory crash, and ships every fix as a plan you approve first.

All services Search logs · level:errorAll levels Last 1 hour Paused
1,296 lines● 7 errors3 warnings10:40 – 10:42 ×
09:4209:5210:0210:1210:2210:32now
7 errors in the last hour · latest 10:42:15postgres restarted at 10:42:10 · exit 137 · nginx 3 errors Ask Agent
Live ↓
10:41:55nginxnginx-110.244.0.9 - "POST /api/login HTTP/1.1" 200 341
10:41:58postgresLOG: duration: 212.418 ms statement: SELECT * FROM shops
10:42:00nginxnginx-010.244.0.12 - "GET /api/coffee-shops/17 HTTP/1.1" 200 1288
10:42:03nginxnginx-0[warn] upstream response is buffered to a temporary file
10:42:06nginxnginx-110.244.0.9 - "GET /api/health HTTP/1.1" 200 15
10:42:10postgresLOG: server process (PID 71) was terminated by signal 9: Killed
10:42:10postgresLOG: terminating any other active server processes
10:42:11nginxnginx-0[error] connect() failed (111: Connection refused) while connecting
10:42:11postgresLOG: database system was not properly shut down; recovering
10:42:12nginxnginx-110.244.0.9 - "POST /api/reviews HTTP/1.1" 502 157
10:42:13postgresFATAL: the database system is starting up
10:42:14nginxnginx-010.244.0.12 - "GET /api/me HTTP/1.1" 502 157
10:42:15postgresLOG: redo done at 0/1A4C0F0 system usage: CPU: user: 0.02 s
10:42:15nginxnginx-110.244.0.9 - "GET /api/health HTTP/1.1" 200 15
10:42:16postgresLOG: database system is ready to accept connections
10:42:16nginxnginx-010.244.0.12 - "GET /api/coffee-shops?page=2 HTTP/1.1" 200 5088
10:42:17nginxnginx-110.244.0.9 - "POST /api/reviews HTTP/1.1" 201 412
10:42:18postgresLOG: checkpoint starting: end-of-recovery immediate wait
10:42:19nginxnginx-010.244.0.12 - "GET /api/me HTTP/1.1" 200 288
10:42:20postgresLOG: connection authorized: user=app database=app SSL enabled
10:42:21nginxnginx-110.244.0.9 - "GET /api/health HTTP/1.1" 200 15
10:42:22nginxnginx-010.244.0.12 - "GET /api/favorites HTTP/1.1" 200 932

Ship your code, wire databases, and route AI models, in one place, with your team and your Ops Agent.

acme-saas ArchitectureJLMKRSShare
Add Redis for sessions and route support chat through the AI Gateway.
Added redis and wired REDIS_URL into web. Support chat now calls Claude through your gateway key, so no provider keys live in the app.
Plan #7Appliedapproved by Jamie
+redisRedis 7.2, 1 GiB
~webREDIS_URL, AI_GATEWAY_URL
~ai-gatewaykey scoped to web
Ask the Ops Agent…
web
acme/saas · main
Live 2 replicas
app.acme.run
postgres
PostgreSQL 16 · replicated
Running
redis
Redis 7.2 · sessions
RunningNew
ai-gateway
Claude · GPT-5 · one key
1.2k calls today
assets
Object storage · 12.4 GB
Public read
Add service

App hosting

Sealos turns local code, HTML decks, GitHub repos and templates into services that run 24/7.

Replaces

Managed databases

PostgreSQL, MySQL, MongoDB and Redis, standalone or replicated. Backed up daily.

Replaces

Object storage & AI Gateway

S3-compatible buckets for your files. One endpoint and one key for OpenAI and Anthropic models.

Replaces

Deploy any project

A local folder, a GitHub repo or an open-source template. Sealos works out the build and the port, so there is nothing to configure.

Local project

Any project on your machine, from a full codebase to a single HTML page.

Describe what you want to deploy…
Select project
Search
Open folder
Work outside a project

GitHub repository

Paste a repo URL. Sealos installs dependencies, builds it and gives you a live link.

GitHub repo
Build and deploy from a repository
Repository URL
github.com/acme/storefront
Branch
main
Deploy

Open-source template

Dify, n8n, WordPress, Open WebUI and 240 more. One click, ready to run.

Template
One-click apps from the library
Search 240+ templates
AllAIToolsBackendDevOps
difyLLM app development platform for agents and RAG
new-apiAI model gateway with multi-provider routing
n8nWorkflow automation with the speed of no-code
wordpressThe open-source CMS behind a third of the web

Agent stays on duty after the deploy

Your Ops Agent reads logs and metrics, proposes fixes, and runs every change as a plan you approve first.

  1. It reads the logs and the metrics together and names the cause, like an out-of-memory crash, before proposing a fix you approve.

  2. Point at any element of a live page and describe the change. It ships as a new revision, with a rollback one click away.

  3. Ask for more space and the volume grows while the database stays online. Backups run daily, and an encrypted tunnel opens on request.

  4. Failed deploys, unhealthy instances, restarts and resources close to their limit surface as hints before your users notice.

Diagnose

It reads the logs and the metrics together and names the cause, like an out-of-memory crash, before proposing a fix you approve.

new-api RestartingLogsMetrics
03:15:52 [warn] pg pool 20/20
03:15:54 [fatal] new-api-0 OOMKilled (exit 137 · 512MiB)
Root causeExit 137 · 96% memory peak
Deploy plan #4Approve 512 MiB → 1024 MiB

Hot updates

Point at any element of a live page and describe the change. It ships as a new revision, with a rollback one click away.

quote-deck.sealos.run/pricing800×600
± Changes +154 −8Hot update #12
<section.pricing-card>Selected
Enterprise Cloud plan$49/mo
8/8 static assets 200 OKRollback

Scale & backups

Ask for more space and the volume grows while the database stays online. Backups run daily, and an encrypted tunnel opens on request.

saas-pgReplicated Live
Storage volume20 GiB40 GiB
Daily 12:00 backup · 14 days✓ Active
127.0.0.1:15432 ⇄ saas-pg.ns-prod.svc:5432
Online resize to 40 GiB, no downtimeTLS tunnel ready

Early warnings

Failed deploys, unhealthy instances, restarts and resources close to their limit surface as hints before your users notice.

Service hints · 7 Live
Deploying updatenew-api
Last deploy failedbilling-worker
Partial instances unhealthy · 1/2auth-ssr
Recently restartedwebhook-relay
Approaching limit · 96%new-api
Rolling out revision #18 · 1/2 pods updated

Everything your app depends on

Databases, storage, model access and your team, provisioned next to the app and wired in from day one.

Databases

PostgreSQL, MySQL, MongoDB or Redis in one click. Connection string injected, backups daily.

postgres
PostgreSQL 16
mysql
MySQL 8.0
mongo
MongoDB 8.0
redis
Redis 7.2

Object storage

S3-compatible buckets for uploads and assets. Drag files in, share a temporary link.

Buckets
3 buckets · 12.4 GB of 20 GB used
assetsPublic
uploadsPrivate
backupsPrivate
Connection
Access key
mz0dmtig
Secret key
••••••••••••••
Endpoint
s3.sealos.run

AI Gateway

One key and one endpoint for GPT-5, Claude, Gemini and DeepSeek. Billed from a single workspace balance.

prod
AI Gateway key
OverviewConnections
Services that call models with this key.
webOpenAI
OPENAI_API_KEY
OPENAI_BASE_URL
workerAnthropic
ANTHROPIC_API_KEY
ANTHROPIC_BASE_URL

Team workspaces

Invite your team into a shared workspace. Everyone works on the same projects with their own credentials.

PersonalOwner
Community
Client
Create new workspace
Invite members
Workspace settings

Common questions

Full-stack codebases, vibe-coded projects, GitHub repositories, interactive HTML pages and client decks, production databases (PostgreSQL, MySQL, MongoDB, Redis), and open-source stacks like WordPress, n8n, Dify, New API and Open WebUI. Every one gets a public HTTPS link.

Hire your dedicated Ops Agent and ship in minutes.