CLI Reference

pip install pie-server

pie run <inferlet> [-- <args>]

# Run standard inferlet
pie run std/text-completion -- --prompt "Hello world"
 
# Run custom inferlet
pie run ./my_inferlet.wasm -- --input "test"
 
# Run with multiple args
pie run my-inferlet -- --prompt "test" --max-tokens 100

pie serve [options]

# Start server
pie serve
 
# Custom port
pie serve --port 9000
 
# Specific model
pie serve --model llama-3.2-1b

pie upload <name> <wasm-file>

pie upload my-inferlet ./target/wasm32-wasi/release/my_inferlet.wasm

pie list

pie build [path]

# Build current directory
pie build
 
# Build specific project
pie build ./my-inferlet

pie init <name> [--lang <language>]

# Rust project (default)
pie init my-inferlet
 
# JavaScript project
pie init my-inferlet --lang js

pie logs [--follow]

pie status

[server]
host = "localhost"
port = 8080
 
[models]
default = "llama-3.2-1b"
cache_dir = "~/.pie/models"
 
[runtime]
max_concurrent = 10
timeout = 300

# Initialize project
pie init my-app
cd my-app
 
# Build inferlet
pie build
 
# Run locally
pie run ./target/wasm32-wasi/release/my_app.wasm -- --prompt "test"
 
# Upload to server
pie upload my-app ./target/wasm32-wasi/release/my_app.wasm

# Start server with specific model
pie serve --host 0.0.0.0 --port 8080 --model llama-3.2-1b
 
# Monitor logs
pie logs --follow