Quick Start
Get your first FiberWise agent running in under 10 minutes. This guide gets you from zero to your first working agent.
5 Minutes Setup
Install and initialize your local environment
First Agent
Create and run your first intelligent agent
Web Interface
Access the full platform through your browser
Prerequisites
Before starting, make sure you have:
- Python 3.8+ installed on your system
- Internet connection for package downloads
- 5 minutes of your time
💡 What You'll Build
A simple text processing agent that can analyze sentiment, summarize text, or generate responses - all running locally on your machine.
Step 1: Set Up FiberWise
FiberWise is a local development platform. Set up your environment by navigating to your Fiberwise installation directory:
# Navigate to your Fiberwise installation
cd path/to/fiberwise
# Install Python dependencies
pip install -r requirements.txt
Verify the setup by checking the directory structure:
# List the main components
ls -la
# You should see: fiber-apps/, fiberwise/, fiberwise-core-web/, etc.
Step 2: Start the Platform
Start the FiberWise web platform:
# Start the web platform (from fiberwise-core-web directory)
cd fiberwise-core-web
python main.py
You'll see output like:
🚀 Initializing FiberWise platform... 📁 Step 0: Storage Configuration Setup 📋 Step 1: Database and Core Setup 🌐 Step 2: Web Components Setup 🎉 FiberWise initialization completed successfully!
Step 3: Create Your First Agent
Create a simple agent that processes text input. Create a new file called my_first_agent.py
:
# my_first_agent.py
async def run_agent(input_data, fiber, llm_service):
"""
A simple text processing agent.
Args:
input_data (dict): Input containing 'text' to process
fiber: FiberWise platform interface
llm_service: LLM provider (optional)
Returns:
dict: Response with processed text and analysis
"""
text = input_data.get('text', 'Hello, World!')
# Simple text analysis
word_count = len(text.split())
char_count = len(text)
# If LLM service is available, get AI analysis
ai_analysis = "No AI provider configured"
if llm_service:
try:
prompt = f"Analyze this text in one sentence: {text}"
response = await llm_service.generate_completion(prompt=prompt)
ai_analysis = response.get('text', 'AI analysis failed')
except Exception as e:
ai_analysis = f"AI analysis error: {str(e)}"
return {
'status': 'success',
'original_text': text,
'word_count': word_count,
'character_count': char_count,
'ai_analysis': ai_analysis,
'agent_name': 'My First Agent',
'message': '🎉 Your first agent is working!'
}
Step 4: Test Agent Instantly (Local Client)
Test your agent immediately using FiberWise's local client. This runs without starting the full platform:
# Run agent directly using local client (no background services needed)
fiber activate my_first_agent.py --input-data '{"text": "FiberWise is amazing!"}'
💡 What's happening here?
The fiber activate
command uses FiberWise's local client to run your agent directly. This is perfect for:
- 🚀 Instant testing - No platform startup required
- 🛠️ Development workflow - Quick iterations and debugging
Learn more about the local client in our Python SDK documentation.
Expected Output:
{
"status": "success",
"original_text": "FiberWise is amazing!",
"word_count": 3,
"character_count": 20,
"ai_analysis": "No AI provider configured",
"agent_name": "My First Agent",
"message": "🎉 Your first agent is working!"
}
Step 5: Start the Platform (See Activity History)
Now let's start the full platform to see your agent activation recorded in the web interface:
# Start the platform in development mode
fiber start --dev
Expected output:
🚀 Development mode enabled (--dev) Database initialization completed successfully 🌐 Starting Vite dev server on port 5556... ✅ FiberWise development server is running! Open your browser to: http://localhost:8000
🔍 What you'll discover:
- Activation History - Your previous agent run is automatically recorded
- Agent Management - Upload, edit, and test agents through the web interface
- Real-time Monitoring - Watch agent activations happen live
- Performance Metrics - Execution time, success rates, and more
Step 6: Explore Both Modes
🚀 Local Client Mode
Best for: Development, testing, automation
- Instant execution
- No background processes
- Perfect for scripts and CI/CD
- Command-line focused
fiber activate agent.py --input-data '{...}'
🌐 Platform Mode
Best for: Production, monitoring, collaboration
- Web-based management
- Activity tracking and history
- Real-time monitoring
- Multi-user environments
fiber start --dev
💡 Recommended Workflow:
- Develop using local client (
fiber activate
) for fast iterations - Test through the web interface for comprehensive validation
- Deploy using platform mode for production monitoring
Step 7: Explore the Web Interface
🏠 Dashboard
View your agents, recent activations, and system status
🤖 Agents
Manage your agents, view their code, and test them interactively
📊 Activations
See the history of agent runs, inputs, outputs, and performance metrics
⚙️ Settings
Configure LLM providers, manage API keys, and adjust platform settings
🎉 What's Next?
Troubleshooting
❌ "fiber: command not found"
Solution: FiberWise not installed or not in PATH
# Reinstall FiberWise
pip install --upgrade fiberwise
# Check if it's in your PATH
which fiber # On Unix/Mac
where fiber # On Windows
❌ "Port 8000 already in use"
Solution: Another service is using port 8000
# Use a different port
fiber start --dev --port 8080
# Or stop the conflicting service
# On Windows: netstat -ano | findstr :8000
# On Unix/Mac: lsof -i :8000
❌ "Web components not found"
Solution: Re-initialize FiberWise
# Force re-initialization
fiber initialize --force
# Or if that fails, try:
fiber initialize --force-clone
❌ Agent activation fails
Solution: Check agent syntax and function signature
- Ensure your agent file has the correct function signature:
async def run_agent(input_data, fiber, llm_service)
- Check for Python syntax errors in your agent file
- Verify the file path is correct
# Test Python syntax
python -m py_compile my_first_agent.py
❌ Web interface not accessible
Solution: Check server startup and firewall
- Verify the server started successfully (check terminal output)
- Try accessing
http://127.0.0.1:8000
instead - Check if firewall is blocking the connection
- Try starting with verbose output:
fiber start --dev --verbose