Maritimemaritime

Cloud hosting for
your AI agents.

Maritime is the easiest way to deploy AI agents to production.
We handle infrastructure, scaling, and routing for $1/month

Works with your stack

Python
Docker
LangChain
CrewAI
OpenAI
AutoGen
Anthropic
Hugging Face
FastAPI
LangGraph
Vercel
Next.js
Python
Docker
LangChain
CrewAI
OpenAI
AutoGen
Anthropic
Hugging Face
FastAPI
LangGraph
Vercel
Next.js

You build your agent. Maritime runs it.

Everything your
agent needs to run.

Your own container, public URL, secrets, metrics, and scaling - all in one place.

Flat $1/month

No per-request fees, no surprise bills.

Built‑in API endpoint

Every agent gets a stable HTTPS URL on deploy. Call it from your app, webhooks, or other agents.

API endpoints

Bring your own stack

OpenClaw, CrewAI, LangGraph, OpenAI Agents, or your own framework. If it runs in a container, it runs on Maritime.

One-click deploys

Connect a GitHub repo and deploy. Your agent is live in seconds - no YAML, no manual wiring.

Built‑in secrets

Store API keys and credentials that are encrypted and injected at runtime via the dashboard.

OPENAI_API_KEY********************
REDIS_URL********************
DATABASE_URL********************
ANTHROPIC_API_KEY********************
GOOGLE_AUTH_SECRET********************

Logs & scaling

Live request logs and basic metrics included. Maritime scales your agent up as traffic grows.

Production-ready in 3 steps.

Push your code
mtm/support-agentPython
mtm/research-botTypeScript
mtm/data-pipelinePython
Push your code
Add secrets
OPENAI_API_KEY=••••••••••••
DATABASE_URL=••••••••••••
WEBHOOK_SECRET=••••••••••••
REDIS_URL=••••••••••••
SENTRY_DSN=••••••••••••
Add secrets
You're live
You're live

Building an AI agent locally is easy.
Deploying it to the cloud is hard.

Unpredictable costs

Per request, per minute, per GB pricing. One traffic spike costs more than a month of dev.

Too much setup

Dockerfile, cloud config, domain, SSL, CI/CD. Days of work before serving one request.

Dockerfile
docker-compose.yml
nginx.conf
deploy.yaml
.env.production
ssl/cert.pem

Built for web apps, not agents

Functions time out. Containers lose state. Not built for AI that thinks for minutes.

agent thinking...4:58
ERROR504 Gateway Timeout
WARNState lost on restart

Built for anyone shipping agents.

Developers

Ship agents without managing infrastructure. Focus on prompts and logic, not DevOps.

1import { OpenClaw } from 'maritime'
2
3const agent = new OpenClaw({
4 version: '3.11',
5 model: 'gpt-5.3,
6 sandbox: false,
7 persistence: true,
8})
9
10await agent.deploy()

Startups

Go from prototype to production in minutes. Scale when you need to, not before.

Researchers & hackathon teams

Deploy demos instantly. Share a live link instead of a localhost screenshot.

Businesses

Run internal AI tools on reliable infrastructure with encrypted secrets and team access.

MemberRoleAgents

Why we're different.

MaritimeMaritime
AWS Lambda
AWS EC2
Google Cloud
DigitalOcean
Railway
Heroku
Fly.io
Render
Monthly cost
$1/mo flat
$18+/mo
API Gateway + NAT extras
$8–15/mo
Pays even when idle
$6–49/mo
Always-on hits $49+
$16/mo
You manage the server
$5–10/mo
Usage-based, ephemeral
$7–25/mo
Add-ons push costs up
$6+/mo
IPv4 + volumes extra
$7+/mo
Database costs extra
Setup time
Minutes
Hours
Hours
Hours
Minutes
Minutes
Minutes
Medium
Minutes
No cold starts
None
Yes
None
Yes
None
None
Yes
None
None
Stateful containers
Yes
No
Yes
No
Yes
No
No
Yes
No
Sleep / wake
Built-in
No
No
No
No
No
No
No
No
Agent-first design
Yes
No
No
No
No
No
No
No
No

Pricing based on running an always-on AI agent (~1 vCPU, 1 GB RAM) as of March 2026. AWS Lambda costs include API Gateway and NAT Gateway for VPC-attached workloads. All prices are approximate monthly on-demand rates in US regions and may vary.

From the blog.

InfrastructureLatest

Traditional Hosting Is Broken for AI Agents

VMs, containers, serverless — none of it was designed for AI agents. Here's why every existing hosting model fails, and what agent-native infrastructure actually looks like.

Mar 11, 20268 min read
Read article
Guides

LangGraph vs CrewAI: Choosing the Right Agent Framework

An honest comparison of LangGraph and CrewAI for building production AI agents — when to use each, their tradeoffs, and how to deploy both.

5 min readRead
Security

Security Best Practices for Production AI Agents

AI agents have unique security requirements — they execute code, call APIs, and handle sensitive data. Here are the practices that actually matter.

6 min readRead
Tutorials

Deploying CrewAI Agents to Production in 5 Minutes

A step-by-step guide to taking your CrewAI multi-agent crew from local development to a live API endpoint with webhooks and monitoring.

4 min readRead
View all articles

Loved by developers.

Alex C.
Alex C.@alex***_dev·Feb 18

Just deployed my CrewAI agent on @maritime_sh in literally 3 minutes. No Dockerfile, no K8s, just… worked? $1/month is insane.

Sarah N.
Sarah N.@sarah***·Mar 2

We moved 4 agents from Lambda to Maritime last week. Cold starts gone. Bill went from $340/mo to $4/mo. Not a typo.

Marco R.
Marco R.@marco***·Feb 24

The sleep/wake architecture is genuinely clever. My agent handles 2k+ requests/day and sleeps between bursts. Paying less than my morning coffee.

Devlin P.
Devlin P.@devlin***·Mar 5

Spent 2 weeks trying to set up K8s for a single demo agent. Tried @maritime_sh on a whim — deployed in 5 minutes. Literally deleted my terraform files.

Emily Z.
Emily Z.@emily***·Feb 28

Maritime's CLI is *chef's kiss*. `maritime deploy` and you're live. Finally something built for AI engineers, not DevOps teams.

Jake R.
Jake R.@jake***·Mar 7

Our hackathon team won because we spent time building instead of debugging Docker networking. 10/10 would recommend @maritime_sh

Simple, transparent pricing.

Start at $1/month. Scale when you need to.

Smart Agent

$1/mo per agent
  • Sleep/wake architecture
  • 1,000 invocations/month
  • Shared gateway
  • Community support
  • Basic logs
Popular

Extended

$5/mo per agent
  • Everything in Smart
  • 10,000 invocations/month
  • Dedicated gateway
  • Webhook triggers
  • Priority logs & analytics

Always-On

$10/mo per agent
  • Everything in Extended
  • Unlimited invocations
  • Persistent connection
  • Priority support
  • Custom domain

Common questions.

Maritime supports any agent that runs in a Docker container — including OpenClaw, CrewAI, LangGraph, AutoGen, and custom Python or Node.js agents. If it runs in Docker, it runs on Maritime.

Each agent gets a flat $1/month base that includes sleep/wake management, a public endpoint, and encrypted secrets. You only pay for compute when your agent is actually awake and processing requests.

Your agent sleeps automatically when idle, using zero compute resources. When a request comes in, Maritime wakes it up in milliseconds — no cold starts, no lost context.

All environment variables are encrypted at rest. Agent containers run in isolated environments. We never access your agent's data or conversation logs.

Join the waitlist and we'll send you an invite within two weeks. Once you're in, you can deploy your first agent in under five minutes from the dashboard or CLI.

Ready to ship your agent?

Join the waitlist and deploy your first AI agent in minutes.
No infrastructure required.

Join Beta