Crewship – Deploy AI agents to production in one command

Name: Crewship – Deploy AI agents to production in one command
Availability: InStock
Author: vhalasi

by vhalasi·Feb 27, 2026·11 points·3 comments

Visit Project View on HN

AI Analysis

●MidShip It

One-command agent deploy, but infrastructure wrappers already exist (LangSmith, Modal, Vercel).

Strengths

•Framework-agnostic: works with CrewAI, LangGraph, and LangGraph.js from day one
•Real-time SSE streaming and instant rollback with version control are solid operational features
•Auto-scaling from zero removes a legitimate DevOps friction point for agent workflows

Weaknesses

•LangSmith already handles agent monitoring, versioning, and deployment — not a differentiated feature set
•Closed-source SaaS; unclear moat over open-source alternatives like LangGraph's own deployment story

Post Description

Hey HN! We built Crewship (https://crewship.dev) because deploying AI agents to production is still unnecessarily painful.

If you've built something with CrewAI, LangGraph, or similar frameworks, you know the drill: it works great locally, then you spend days figuring out infrastructure, scaling, monitoring, and artifact management just to get it running for real users.

Crewship handles all of that. You add a crewship.toml to your project, run `crewship deploy`, and your agents are live in seconds. It's framework-agnostic — we currently support CrewAI, LangGraph, and LangGraph.js, with more coming.

What you get: - One-command deploy (no Docker/K8s config needed) - Real-time streaming of agent actions via SSE - Automatic artifact collection (every file/report your agents produce) - Auto-scaling from zero to thousands of concurrent runs - Version control with instant rollback - Secrets management

We're a small team in Switzerland, and we've been using this ourselves for months. Free tier available — would love your feedback.

Docs: https://docs.crewship.dev Quickstart: https://docs.crewship.dev/quickstart