Skip to main content

Flowise Templates

Agent flow templates for visual AI development

The resources/flowise/ folder contains pre-built agent flows that demonstrate common AI patterns and integrations with Seiling services.

Available Templates​

Seiling Flowise Agentflow Demo​

  • File: Seiling Flowise Agentflow Demo Agents.json
  • Purpose: Complete agent building demonstration
  • Features: Multi-service integration, common AI patterns
  • Best for: Understanding Flowise capabilities

Template Features​

Core Components​

  • LLM Integration: OpenAI, Anthropic, Google AI support
  • Memory Management: Conversation context preservation
  • Tool Integration: Custom tools and APIs
  • Vector Storage: Document embeddings and search
  • Chain Orchestration: Multi-step reasoning flows

Seiling Integration​

  • Sei MCP Tools: Blockchain operations
  • Database Connections: PostgreSQL, Redis access
  • Service APIs: n8n, OpenWebUI integration
  • Custom Functions: Seiling-specific operations

Import Instructions​

Step 1: Access Flowise​

# Open Flowise interface
http://localhost:5003

# Login credentials
Username: admin
Password: seiling123

Step 2: Import Template​

  1. Chatflows → Load Chatflow
  2. Upload file from resources/flowise/
  3. Import and Save

Step 3: Configure​

  1. API Keys: Add OpenAI, other AI service keys
  2. Service URLs: Update if using custom ports
  3. Database Connections: Configure if needed
  4. Test Flow: Run test queries

Agent Patterns​

Basic Chatbot​

User Input → LLM Processing → Response

RAG (Retrieval Augmented Generation)​

User Query → Vector Search → Context + LLM → Response

Tool-Enabled Agent​

User Input → Intent Detection → Tool Selection → Tool Execution → Response

Multi-Step Reasoning​

User Query → Planning → Sub-tasks → Tool Calls → Synthesis → Response

Configuration​

Required Settings​

  • OpenAI API Key: For LLM processing
  • Vector Database: Qdrant connection
  • Memory Store: Redis for conversation context

Optional Integrations​

  • Sei MCP Server: Blockchain tool access
  • PostgreSQL: Persistent data storage
  • Custom APIs: External service integration

Common Use Cases​

DeFi Assistant​

Flow: Query → Market Data → Portfolio Analysis → Recommendations Tools: Sei MCP, Price APIs, Portfolio tracker

Trading Bot​

Flow: Market Signal → Analysis → Risk Assessment → Trade Execution Tools: Sei MCP, DEX APIs, Risk models

Research Agent​

Flow: Topic → Data Gathering → Analysis → Report Generation Tools: Web scraping, Document processing, Summary generation

Portfolio Manager​

Flow: Portfolio Query → Analysis → Optimization → Action Plan Tools: Sei MCP, Analytics, Rebalancing logic

Customization​

Adding New Tools​

  1. Tools → Custom Tool
  2. Configure API: Endpoint, parameters
  3. Add to Flow: Connect to chain
  4. Test Integration: Verify functionality

Memory Configuration​

// Memory settings
{
  "memoryType": "redis",
  "redisUrl": "redis://seiling-redis:6379",
  "sessionTTL": 3600
}

Vector Store Setup​

// Qdrant configuration
{
  "url": "http://seiling-qdrant:6333",
  "collectionName": "seiling_docs",
  "dimension": 1536
}

Advanced Features​

Chain Composition​

  • Sequential Chains: Step-by-step processing
  • Parallel Chains: Concurrent operations
  • Conditional Chains: Decision-based routing
  • Loop Chains: Iterative processing

Context Management​

  • Conversation Memory: Multi-turn dialogs
  • Session Storage: User-specific data
  • Document Context: RAG implementation
  • Tool Memory: Action history

Integration Patterns​

  • API Calls: External service integration
  • Database Queries: Data retrieval and storage
  • File Processing: Document analysis
  • Real-time Data: Live API feeds

Best Practices​

Performance​

  • Optimize Chains: Minimize unnecessary steps
  • Cache Results: Store frequent queries
  • Batch Operations: Group similar tasks
  • Stream Responses: Real-time output

Security​

  • API Key Management: Secure credential storage
  • Input Validation: Sanitize user inputs
  • Rate Limiting: Prevent abuse
  • Access Control: User permissions

Reliability​

  • Error Handling: Graceful failure management
  • Retry Logic: Automatic recovery
  • Fallback Options: Alternative paths
  • Health Monitoring: System status checks

Troubleshooting​

Import Issues​

  • File Format: Ensure valid JSON
  • Version Compatibility: Check Flowise version
  • Dependencies: Verify required nodes available

Runtime Errors​

  • API Limits: Check rate limiting
  • Service Connectivity: Verify endpoints
  • Memory Issues: Check Redis connection
  • Tool Failures: Debug individual tools

Common Fixes​

# Restart Flowise
docker restart seiling-flowise

# Check logs
docker logs seiling-flowise --tail=50

# Verify dependencies
bash scripts/bootstrap/health_check.sh quick

Creating New Templates​

Template Design​

  1. Define Purpose: Clear use case
  2. Plan Flow: Map user journey
  3. Select Tools: Required integrations
  4. Build Incrementally: Start simple, add complexity
  5. Test Thoroughly: All execution paths

Export Template​

  1. Chatflows → Export
  2. Save JSON: Add to resources folder
  3. Document Usage: Add to guide
  4. Share Template: Contribute to community

Flowise templates provide visual AI development patterns. Start with the demo template and customize for your specific needs.