Techies

Techies: A no-code platform for orchestrating collaborative AI agents through YAML configuration.

---

What is Techies?

Techies is a less-code to no-code agent orchestration platform featuring a modular architecture that enables multiple AI agents to collaborate on complex workflows through YAML-driven configuration. The framework includes an extensible tooling system, making it accessible for users who want to orchestrate AI agent interactions without extensive coding knowledge. Techies functions as a laboratory environment where users can experiment with different agent configurations, task assignments, and collaborative workflows to achieve sophisticated outputs.

Developed by the KD Research organization, led by Kaidong Hu.

---

Installation

# Install latest version
pip install git+https://github.com./kd-research/Techies.git

# For specific version (uncomment and modify)
# pip install git+https://github.com./kd-research/[email protected]

This installs:

• techies — comprehensive CLI for all Techies functionality

---

Environment Setup

Set the following environment variables before running Techies:

# Required for LLMs (default: OpenAI)
export OPENAI_API_KEY=<your-api-key>

# Optional (AgentOps monitoring/debugging)
# Note: agentops is optional and requires manual installation:
# pip install agentops
export AGENTOPS_API_KEY=<your-api-key>

# Optional (audio support)
export FREESOUND_CLIENT_API_KEY=<your-api-key>

# Recommended: Specify your model and provider
export MODEL=openai/gpt-4o

Visit LiteLLM's model list for supported providers.

If you're using another LLM provider (e.g. Groq, Anthropic, Mistral), you'll need to export its matching API key too.

---

CLI Tools

techies: CLI Interface

The powerful CLI interface for running and managing crews:

# Explore available components
techies list crews
techies list agents
techies list tasks

# Run and introduce crews
techies run hierarchy_crew_v2 --game tictactoe
techies introduce hierarchy_crew_v2

Techies CLI also supports:

• scaffold — create new custom crews
• dump — extract and modify existing crews
• TECHIES_RUNTIME — define custom runtime paths

techies scaffold my_crew
techies dump hierarchy_crew_v2
techies run mycrew

Note: Legacy commands like list_crews, list_agents are still available but will be removed in v2.0.

---

Documentation

• Getting Started with Techies
• Running Predefined Crews
• Understand Crew Configurations
• Modifying Existing Crews
• Create Your Own Crew
• Create Your Own Tool
• Using Callbacks in Tasks
• Validating Your Configurations

---

Development

Clone & Install Locally

git clone https://github.com./kd-research/Techies.git
cd Techies
pip install -e .

Requires Python 3.12.4+

---

Best Practices

• Use techies scaffold to start your own crew
• Define a clean working directory for each run
• Use TECHIES_RUNTIME to load external crew folders
• Set MODEL and relevant API keys for your LLM provider

---

License & Credits

Techies is open-source and licensed under the GPLv3 License.
Created and maintained by Kaidong Hu at KD Research.