Prefy is a comprehensive AI platform skill providing smart model routing, automated VPS management, and real-world actions like AI-driven phone calls.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install prefy
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install prefy using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Prefy?
Prefy serves as a versatile backbone for developers using Openclaw Skills, offering a unified Conductor API that routes requests across 19+ industry-leading models including GPT-4o, Claude 3.5 Sonnet, and DeepSeek. It functions as a drop-in OpenAI-compatible replacement, allowing agents to switch between speed-optimized and quality-focused strategies seamlessly.
Beyond text generation, Prefy bridges the gap between digital intelligence and physical infrastructure. It enables automated VPS server provisioning, real-time web search capabilities via Tavily, and even automated phone reservations through its AutoCall feature. This makes it an essential tool for building sophisticated agents that require both high-level reasoning and the ability to manage technical environments or execute real-world tasks.
Prefy Use Cases
- Implementing cost-effective AI routing by automatically selecting the fastest or highest-quality model for a specific task.
- Provisioning and controlling dedicated VPS servers through API-driven commands like shell execution and Ollama management.
- Automating restaurant and hotel bookings using AI-powered phone calls with natural language processing.
- Enhancing agent knowledge with real-time web intelligence and dynamic image generation using FLUX models.
How Prefy Works
- Authentication is established using a Prefy API key (pc_ prefix) or a Supabase JWT token depending on the target endpoint.
- The agent sends requests to the Conductor API, which applies specific strategies like :fast or :quality to route the query to the optimal LLM.
- For infrastructure requirements, the skill interacts with the Server Management API to deploy Hetzner-backed VPS instances or execute agent-level commands.
- Real-world tasks such as phone calls are handled by the AutoCall endpoint, which leverages Vapi and ElevenLabs to conduct human-like conversations for bookings.
- Web search and image generation requests are processed through dedicated endpoints that synthesize data from Tavily and Together AI.
Prefy Setup
To begin using this skill within the Openclaw Skills ecosystem, set your authentication headers and use the following base configuration:
# Export your Prefy API key
export PREFY_API_KEY="your_pc_key_here"
# Test the Conductor API connection
curl https://prefy.com/api/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $PREFY_API_KEY" \
-d '{"model":"conductor:fast","messages":[{"role":"user","content":"Hello"}],"stream":true}'
Prefy Data Schema & Taxonomy
The skill organizes data across several specialized API structures to ensure high interoperability within the Openclaw Skills framework:
| Component | Data Format | Key Fields |
|---|---|---|
| Conductor API | OpenAI Compatible | model, messages, stream, strategy |
| Web Search | JSON | query, max_results, snippets |
| Server Mgmt | JSON | serverId, command (stats, shell, ollama) |
| AutoCall | JSON | phone, venue_name, task, language |
| Image Gen | JSON | prompt, model (flux-schnell), base64 output |
Prefy Advanced Features
- Smart Strategy Routing: Append :fast, :quality, :creative, or :private to the conductor model to instantly change routing behavior.
- Server Agent Integration: Execute shell commands, manage cron jobs, and run local LLMs via Ollama directly on provisioned VPS instances.
- Intelligent Intent Detection: The Agent API automatically classifies user messages to trigger search, flight booking, or phone call workflows.
- Multi-Model Versatility: Access a wide range of models including DeepSeek-R1 for chain-of-thought reasoning and Groq-Llama for speed-critical applications.
SKILL.md
Loading
Related Openclaw Skills
Mermaid Workflow Skill
A technical automation suite for generating, rendering, and embedding Mermaid diagrams into Markdown documentation via a structured script-based workflow.
Markdown to Word Converter
A Python-powered utility to transform Markdown content into formatted Word documents with support for templates and complex elements.
Mulch Self Improver
A persistent memory layer that enables AI agents to record learnings and avoid repeating mistakes across coding sessions.
FACEPALM
An automated troubleshooting tool that uses Codex 5.3 to analyze console logs and chat history for instant issue resolution.
RunRelay
A comprehensive travel API that enables AI agents to search and book over 300 airlines, including exclusive low-cost carrier coverage.
HiFi Advisor
A professional audio consultant agent for matching gear, optimizing room acoustics, and analyzing second-hand market pricing.
Featured*
