Daytona

Daytona is secure, elastic infrastructure for running AI generated code. Developers and agents create isolated sandboxes through an SDK, execute code and commands inside them, and delete them when finished, so untrusted output from a model never touches the host infrastructure. Daytona states that sandbox creation takes under 90 milliseconds from code to execution. The platform is aimed at teams building AI agents, coding assistants and evaluation systems that need many concurrent, stateful environments. SDKs are available for Python, TypeScript, Ruby, Go and Java. Programmatic control covers process execution with real time output streaming, file system operations with full create, read, update and delete support and granular permissions, native Git operations with secure credential handling, and built in language server support for completion and analysis across languages. Sandboxes run isolated from one another and can be parallelised heavily for concurrent workflows, and Windows sandboxes and GPU machines are available. Pricing is pay as you go and billed per second, with $200 in free compute included and no credit card required to start. Compute costs $0.0504 per vCPU hour, memory $0.0162 per GiB hour and storage $0.000108 per GiB hour after the first 5 GiB free. Windows sandboxes cost $0.0858 per vCPU hour. GPU rates run from $0.57 per hour for an Nvidia RTX 4090 preemptible instance to $6.25 per hour for an on demand Nvidia B200 or B300, with H100 at $2.27 preemptible or $3.95 on demand, and H200 at $2.61 or $4.54. Larger commitments are handled through the sales team.