Quick Start
This guide walks you through installing the RunVero CLI, executing your first sandboxed WASI 0.2 tool component, and inspecting the live resource breakdown.
1. Installation
Section titled “1. Installation”Quick Install (macOS & Linux)
Section titled “Quick Install (macOS & Linux)”Install the pre-built CLI binary directly to your path:
curl -sSL https://runvero.dev/install.sh | shInstall via Cargo
Section titled “Install via Cargo”If you have the Rust toolchain installed:
cargo install runvero-cliVerify Installation
Section titled “Verify Installation”runvero --helpOutput:
runvero 0.1.0Deterministic WebAssembly control plane host for AI agent tools
USAGE: runvero <SUBCOMMAND>
SUBCOMMANDS: run Execute a WASI 0.2 component tool with strict budgets serve Start the HTTP and Model Context Protocol (MCP) gateway cache Manage the verified OCI content cache help Print this message or the help of the given subcommand(s)2. Execute Your First Sandboxed Tool
Section titled “2. Execute Your First Sandboxed Tool”RunVero executes WebAssembly components targeting wasm32-wasip2.
Execute a mock tool component with sample JSON input:
runvero run target/wasm32-wasip2/release/mock_tool.wasm --input '{"query": "calculate_risk"}'Live Resource Breakdown
Section titled “Live Resource Breakdown”By default, runvero run prints execution telemetry directly beneath the tool result:
✓ Execution succeeded in 0.48msMemory Allocated: 1.06 MiB (Peak: 1.06 MiB / Cap: 16.00 MiB)Fuel Consumed: 14,821 units (Remaining: 9,985,179 units)Security Policy: Zero-Trust WASI 0.2 Sandbox VerifiedOutput: {"risk_score": 0.12, "classification": "low"}3. Useful Execution Flags
Section titled “3. Useful Execution Flags”Output Raw Result Only (For Shell Pipelines)
Section titled “Output Raw Result Only (For Shell Pipelines)”runvero run my_tool.wasm --input '{"data": 42}' --rawEmit Machine-Readable JSON Telemetry
Section titled “Emit Machine-Readable JSON Telemetry”runvero run my_tool.wasm --input '{"data": 42}' --breakdown-jsonOutput:
{ "success": true, "output": "{\"risk_score\": 0.12}", "fuel_consumed": 14821, "peak_memory_bytes": 1114112, "latency_micros": 482}4. Run the Model Context Protocol (MCP) Server
Section titled “4. Run the Model Context Protocol (MCP) Server”You can instantly expose your local tools to Claude Desktop, Cursor, or any MCP client:
runvero serve --host 127.0.0.1 --port 8080 --tools-dir ./toolsNow, point your Claude Desktop or Cursor configuration to http://127.0.0.1:8080/mcp!