
Superserve
β Text source-checked Nov 19, 2026Virtual machines for long-running agents
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked Pricing captured How we rate
In short
Superserve offers isolated microVMs for long-running agents with persistent state, allowing them to pause and resume complex tasks across days or weeks.
Superserve provides sandboxes where agents build and run complex applications, with persistent storage and programmable networking. Superserve is designed for agents that need machines that remember state across tasks. Superserve allows agents to pause, return to, and branch from machines as work evolves.
The vendor says that every Superserve sandbox is its own microVM with its own kernel. Superserve sandboxes do not share kernels or cross-tenant state. When a Superserve sandbox is paused, the whole machine, including files, installed packages, and state, is kept so the agent resumes where it left off. Superserve uses a credentials broker to add API keys to requests outside the machine, only for hosts the user allows.
Superserve supports running Claude Code, Codex, or custom agent loops inside the machine, controlled via TypeScript, Python, MCP, CLI, or API.
Superserve can be deployed in a user's own AWS, GCP, or Azure account.
The vendor says that Superserve Developer and Pro plans have no fixed session time limit, allowing agents to work for days or weeks.
According to Superserve: Superserve provides sandboxes where agents build and run complex applications, With persistent storage and programmable networking. Its pages list this strength: Each sandbox runs as a microVM with its own kernel.
β The AI Tool Bible editorial team
Pros
- β Each sandbox runs as a microVM with its own kernel.
- β Pausing a sandbox preserves the entire machine state, including files and packages.
- β A credentials broker manages API keys outside the machine for allowed hosts.
- β Superserve supports deployment in the user's own cloud VPC.
- β The SDKs, CLI, and MCP server are open-source under the Apache-2.0 license.
Pricing (captured )
| Tier | Monthly | Key features |
|---|---|---|
| Developer | $20 / month |
|
| Pro | $250 / month |
|
| Enterprise | Custom |
|
Prices change; we show the grid as captured on Nov 11, 2026. Check current pricing on the vendor's site β
Facts
- Pricing
- freemium (not source-checked)
- API available
- Yes (not source-checked)
- Free trial
- Yes
- Open source
- Yes
- Hacker News
- 4 stories Β· top 10 points Β· last Jul 21, 2026 Β· checked Nov 21, 2026
- Earliest Wayback capture
- 2026-02-06 (earliest archived capture of the site, not a launch date)
Where this tool sits
Use cases
Frequently asked
- How much does Superserve cost?
- Superserve uses a freemium pricing model. Specific plan details are not documented in the provided information, but Developer and Pro plans are mentioned as having no fixed session time limits.
- Can I run Claude Code or Codex on Superserve?
- Yes, Superserve supports running Claude Code, Codex, or custom agent loops inside the machine. These can be controlled via TypeScript, Python, MCP, CLI, or API.
- Is my data isolated from other users?
- Yes, every Superserve sandbox is its own microVM with its own kernel. Sandboxes do not share kernels or cross-tenant state, ensuring isolation between users.
- Can I deploy Superserve in my own cloud account?
- Yes, Superserve can be deployed in a user's own AWS, GCP, or Azure account, allowing for self-hosted infrastructure management.
- How does Superserve handle API keys and security?
- Superserve uses a credentials broker to add API keys to requests outside the machine. This is only done for hosts that the user explicitly allows.
- What happens when an agent pauses a task?
- When paused, the whole machine, including files, installed packages, and state, is kept. The agent can resume exactly where it left off, supporting long-running workflows.