The Nonhumans CLI is the fastest way to go from zero to a running agent. Use it to scaffold new projects, push deployments, tail logs, inspect your agent’s inbox and wallet, and manage secrets — all without leaving your terminal. It pairs with both the TypeScript and Python SDKs and works equally well in local development and CI pipelines.Documentation Index
Fetch the complete documentation index at: https://docs.nonhumans.ai/llms.txt
Use this file to discover all available pages before exploring further.
Installation
Authentication
Before running any command, authenticate the CLI with your Nonhumans account:~/.nonhumans/credentials and all subsequent commands use it automatically.
In headless environments (CI, Docker), pass your API key via the
--api-key flag or the NONHUMANS_KEY environment variable instead of running login.Command reference
nonhumans init [name]
Scaffold a new agent project in the current directory (or a named subdirectory). Creates a project structure with a starter agent file, a nonhumans.json config, and an example .env.
nonhumans deploy
Build and push your agent to always-on compute. The CLI reads your nonhumans.json for the entry point and runtime, then streams build output to your terminal.
nonhumans status
Check the current runtime status of your deployed agent — whether it is running, sleeping, or in an error state.
nonhumans logs
Tail live log output from your running agent. Press Ctrl+C to stop streaming.
nonhumans email
Inspect your agent’s email inbox from the terminal. Lists recent messages with sender, subject, and timestamp.
nonhumans wallet
Check balances and send payments without opening a dashboard.
nonhumans memory
Search and inspect everything stored in your agent’s vector memory.
nonhumans vault set/get
Store and retrieve secrets without touching the dashboard.
nonhumans agents list
List every agent associated with your account, along with their handle, status, and creation date.
Global flags
Every command accepts these flags:| Flag | Description |
|---|---|
--api-key <key> | Override the stored credential with a specific API key. |
--output json | Print output as JSON — ideal for scripting and piping into jq. |
--help | Print usage information for any command. |
From init to deploy — step by step
The following walkthrough takes you from a blank directory to a live, always-on agent in under five minutes.
Scaffold your project
Create a new agent project. Replace The scaffolded directory looks like this:
my-agent with your desired handle.Write your agent logic
Open
agent.ts and build your agent using the SDK. The scaffold includes a minimal example to get you started.Deploy to always-on compute
Push your agent. The CLI bundles your code, uploads it, and starts the runtime.You will see streaming build output. A successful deploy prints your agent’s public URL:
Next steps
TypeScript SDK
Explore all SDK modules — email, wallet, memory, compute, and more.
Python SDK
Build agents in Python with sync and async support.