The problem I kept running into: most sandbox providers kill your agent after 24 hours. If you're running something autonomous that needs to work for days — refactoring a codebase, running tests in a loop — you're constantly fighting timeouts and rebuilding state.
Superserve lets you snapshot a running VM at any point, fork it into parallel branches, and resume exactly where you left off. Each agent gets its own VM, no shared kernels. There's also a credentials broker so agents never touch raw API keys directly, and egress is locked to URLs you specify.
You bring your own agent (Claude Code, Codex, etc.). We handle the durable compute underneath. TypeScript and Python SDKs, MCP support, open-source and self-hostable.
Would love to hear feedback, especially from anyone building long-running agents.