Zeabur review

Zeabur is a hosted deployment and infrastructure platform, not a library or self-hosted server. It sits between a code repository or local project and the compute that runs the application. The homepage says its agent processes code, detects the language, analyzes the framework, and handles deployments.

The platform also covers servers and clusters, domains, DNS, email, and ready-made application templates. Its agent Skills turn natural-language prompts into infrastructure actions, including deploying services, renting servers, and managing configurations. That places Zeabur closer to a managed DevOps control plane than a single-purpose frontend host.

The architecture is attractive when a team wants one place to handle deployment and adjacent infrastructure rather than assemble separate tools. But reported shutdown notices change the decision: availability and the ability to recover projects elsewhere matter more than the breadth of the control plane. Confirm the service’s status before relying on it.

The integration surface described on the homepage is the Zeabur Agent: developers can use natural-language prompts to deploy services, rent servers, and manage configurations. The homepage also says users can deploy a local project or a template, with the agent detecting its framework. For email, Zeabur says it can configure SMTP credentials automatically.

The homepage does not name a REST API, webhooks, CLI, plugins, SDKs, or supported programming languages. It describes code-language and framework detection, but does not enumerate the languages or frameworks. Confirm API access, SDK support, CLI availability, and the deployment workflow in Zeabur’s vendor documentation before designing automation around them.

The prompt-based path may reduce setup work for a first deployment. It is not a substitute for checking how configuration, secrets, build output, and operational events fit into an existing delivery pipeline. Ask the vendor how those interfaces work and whether they remain available during migration or service closure.

The Free plan costs $0 per month and includes one manageable server, a 2C4G build machine, 48-hour log retention, and a 50 MB single-file upload limit. The pricing lists a 14-day free trial. Dev costs $5 per month for three manageable servers, a 2C4G build machine, seven-day log retention, 30 advanced-model messages, and 250 MB uploads.

Pro is $19 per month and includes 10 servers, a 4C8G build machine, 30-day logs, 200 advanced-model messages, and 1 GiB uploads. Team costs $79 per month, includes unlimited manageable servers, 90-day logs, three Pro-level seats, and 600 advanced-model messages. Additional Team seats cost $24 per seat per month. Enterprise pricing is custom.

The pricing page lists monthly prices and no annual billing or discount. Zeabur describes fixed plans, but its email service is usage-based. The listed plans do not establish runtime compute charges or usage overages; ask how server rental, email usage, and limits are billed before production use.

Zeabur’s homepage does not identify an open-source license or describe a self-hosted edition. It presents Zeabur as a platform for deploying projects and managing infrastructure, so treat the service as a hosted dependency unless the vendor confirms that its control plane can be self-hosted.

The homepage also does not describe a data-export format or a complete migration path. That leaves key exit questions open: whether deployment configuration, environment variables, logs, domains, email settings, and persistent data can be exported, and which parts must be recreated on another platform.

Before committing, ask for a documented export procedure and test whether the application can be redeployed from its repository and independent backups. One Reddit commenter warned that being unable to redeploy from a repo alone could make a platform shutdown the real blast radius. Do not assume the control plane or managed services transfer with your code.

The plan limits matter operationally: Free has 48-hour logs and a 50 MB single-file upload limit, while log retention rises to 90 days on Team. Advanced-model messages are capped by plan, and the homepage describes email as usage-based. The vendor should clarify runtime compute charges, service-level commitments, backup and recovery procedures, and the support response available on each plan.

There is a larger continuity concern. A sysadmin Reddit post describes an August 2026 breach involving a vulnerable customer application and exposed credentials, databases, and customer information, followed by incomplete remediation. These are claims in a Reddit post, not independently verified incident details. In another thread, users say they received shutdown notices. Confirm current availability directly with Zeabur before deploying.

For any existing workload, keep source code, configuration, secrets, and persistent data recoverable outside the platform. A SillyTavern user described taking about four hours to work through an export with an AI assistant. That is an individual report, not a general migration estimate, but it illustrates why an exit plan matters.