cyberian

A Python CLI wrapper for agentapi that enables automated agent pipelines and workflows.

What is cyberian?

cyberian provides a comprehensive command-line interface for orchestrating AI agents through agentapi. It enables you to:

• Send messages to AI agents via agentapi with synchronous or asynchronous modes
• Manage agentapi servers - start, stop, and list running servers
• Run server farms - manage multiple agent servers simultaneously
• Execute complex workflows from YAML definitions with recursive task trees
• Use template-based instructions with Jinja2 for dynamic, parameterized workflows
• Call external providers directly in workflows for deterministic operations

Quick Start

# Install
pip install cyberian

# Start an agentapi server
cyberian server start claude --skip-permissions --port 3284 --dir /tmp/workdir

# Send a message
cyberian message "Write a hello world function in Python"

# Check status
cyberian status

# View in chat interface
open http://localhost:3284/chat

Key Features

Simple Message Sending

# Fire-and-forget message
cyberian message "Your prompt here"

# Synchronous mode - wait for response
cyberian message "What is 2+2?" --sync

Server Farm Management

Start and manage multiple agent servers from a single configuration:

cyberian farm start my-farm.yaml

base_port: 4000
servers:
  - name: worker1
    agent_type: claude
    directory: /tmp/worker1
    skip_permissions: true
  - name: worker2
    agent_type: claude
    directory: /tmp/worker2

Workflow Automation

Execute complex multi-step workflows with recursive task trees:

cyberian run deep-research.yaml --param query="quantum computing"

name: deep-research
description: Iteratively performs deep research on a topic

params:
  query:
    range: string
    required: true

subtasks:
  initial_search:
    instructions: |
      Perform deep research on {{query}}.
      Write a comprehensive research plan.

  iterate:
    instructions: |
      Continue researching {{query}}.
      Explore new angles and perspectives.

    loop_until:
      status: NO_MORE_RESEARCH
      message: |
        If exhausted, yield status: NO_MORE_RESEARCH

Installation

# Basic installation
pip install cyberian

# With provider support for external API calls
pip install cyberian[providers]

# Or use without installation
uvx cyberian

Where to Go Next

• Tutorials

  ---

  New to cyberian? Start here to learn the basics.

  Getting Started

• How-To Guides

  ---

  Task-oriented guides for specific problems.

  Browse Guides

• Explanation

  ---

  Understand how cyberian works under the hood.

  Learn Concepts

• Reference

  ---

  Complete reference for all commands and schemas.

  CLI Reference

Integration with Claude Code

A Claude Code skill is available for multi-agent orchestration:

/plugin marketplace add cyberian-skills

See Claude Code Skill for details.

Project Links

• Documentation: https://monarch-initiative.github.io/cyberian
• Source Code: https://github.com/monarch-initiative/cyberian
• Issue Tracker: https://github.com/monarch-initiative/cyberian/issues
• agentapi: https://github.com/coder/agentapi

Technology Stack

• Python 3.10+ with uv for dependency management
• Typer for CLI interface
• httpx for HTTP client interactions
• Pydantic for data modeling and validation
• Jinja2 for template rendering in workflows
• pytest for comprehensive testing