Maritime

What is Maritime?

Maritime is cloud hosting for AI agents that deploys them to production with a live HTTPS API endpoint. Instead of managing infrastructure yourself, you provide an agent in a container and Maritime runs it with routing, scaling, logs, and secret handling.

Its core purpose is to make it practical to move an agent from local development to a reachable service—so your application, webhooks, or other agents can call it via a stable endpoint.

Key Features

• Deploy agents as containers: If your agent runs in a container, it can run on Maritime—supporting frameworks like OpenClaw, CrewAI, LangGraph, and other agent setups.
• Built-in stable HTTPS API endpoint: Each deployed agent gets a public URL on deploy so you can integrate it into apps and workflows.
• Bring your own stack: Maritime is designed to work with your agent code and framework choices as long as it runs in a container.
• One-click deploy from GitHub: Connect a GitHub repo and deploy without manually wiring YAML or doing extensive cloud configuration.
• Encrypted secrets injected at runtime: Store credentials in the Maritime dashboard; secrets are encrypted and injected into the running agent environment.
• Live request logs and basic metrics: View request logs and basic metrics, and use them alongside scaling behavior as traffic changes.

How to Use Maritime

1. Create your agent container and ensure it can run in a Docker-style container environment.
2. Connect a GitHub repo that contains your agent code and deployment configuration.
3. Deploy from Maritime to get a public HTTPS URL for your agent.
4. Add secrets in the Maritime dashboard (for example, API keys and other credentials). Maritime injects them at runtime.
5. Call your agent from your app, from webhooks, or from other agents using the endpoint provided after deploy.

Use Cases

• Build an app-backed support agent: Deploy a support agent so a web app can call it via a stable HTTPS endpoint instead of running locally.
• Run a research bot for scheduled or on-demand workflows: Host an agent that can be triggered by your systems through API calls or webhooks.
• Deploy a data pipeline agent: Expose an agent as an API service to orchestrate or trigger data processing steps from other services.
• Move multi-agent workflows to production: Host agents built with frameworks such as CrewAI or LangGraph so you can route traffic to a live service and observe requests.
• Team demos and prototypes with a shareable link: Replace “localhost screenshots” with a live link by deploying quickly to a public endpoint.

FAQ

• Does Maritime support my agent framework? Maritime states it supports bringing your own stack and that if it runs in a container, it runs on Maritime. It also lists example integrations/frameworks like OpenClaw, CrewAI, LangGraph, and agent setups using OpenAI Agents.

• How does Maritime make my agent reachable? On deploy, Maritime provides a stable HTTPS URL for each agent.

• How are API keys and other credentials handled? Maritime offers a secrets dashboard where secrets are encrypted and injected at runtime.

• What operational visibility do I get? Maritime includes live request logs and basic metrics.

• How do I deploy? The site describes connecting a GitHub repo and using one-click deploy, rather than manual YAML/cloud wiring.

Alternatives

• Traditional server hosting (VMs/containers): Run the agent yourself on infrastructure and configure routing, SSL, and scaling. This shifts operational work to you rather than providing an agent-first deployment flow.
• Serverless functions: Package logic for function-style execution. This can be a better fit for short-lived tasks but may require more attention to constraints like cold starts and execution/time limits.
• Managed workflow/orchestration platforms: Use tools that orchestrate multi-step AI workflows but may not provide the same agent container hosting model and live endpoint experience described by Maritime.