Ephemeral data sandbox for AI workflows with guardrails and security
Run Python code securely in a sandbox environment via MCP tool calls.
Key Features: Secure sandboxed Python code execution Integration with MCP tool call interface Supports dynamic code evaluation within agent workflows Give an agent a real Linux VM: run commands, move files, expose a preview URL, destroy it.
LATAM fiat ramps on Stellar for AI agents: live quotes, sandbox orders, PIX/SPEI simulation, docs
Control Google Colab notebooks, run cells, manage cell contents, and assign T4, L4, or A100 GPUs through MCP.
Key Features: Pre-registers all nine MCP tools at startup Connects to Colab through a browser Reads cells, IDs, contents, and outputs Run multiple isolated containerized environments for coding agents with git branch isolation and terminal access.
Key Features: Isolated container environments per agent with separate git branches Real-time visibility into commands and logs executed by agents Terminal access to agent environments for manual intervention Run LLM-generated JavaScript or Python code securely with YepCode's sandboxed MCP server supporting JSON Schema tool parameters.
Key Features: Polyglot support for JavaScript and Python tool implementations Secure sandboxed code execution environment Tool parameter definitions via JSON Schema Sandbox workspace tools: search, file read, DB queries, integrations. Returns synthetic data.
E-signatures for agents: mint a sandbox key, send PDFs, track status, download the sealed result.
Execute untrusted Python/JavaScript code in isolated WebAssembly sandboxes with resource limits and retries.
Key Features: Isolated execution in WebAssembly sandboxes Configurable CPU (compute) allocation levels Memory and timeout limits per task Proves AI-generated Python does what you asked: lint, types, security, sandbox run, exact fixes.
Statically-typed JSON AST programming language with verification and WASM compilation, exposed via 19 MCP tools.
Key Features: JSON AST as canonical program format with no parser Structured, typed JSON errors for agent self-repair Static type system with effect tracking (pure, reads, writes, io, fails)