🏔️ Snow-Flow

Multi-LLM ServiceNow Development Platform

Develop ServiceNow through conversation • 411 Tools • 2 MCP Servers • 75+ LLM Providers • Works with ANY AI coding assistant

What is Snow-Flow?

Snow-Flow transforms ServiceNow development into a conversational experience. Instead of clicking through ServiceNow's web interface, you develop through natural conversation with your AI coding assistant using any LLM provider you prefer.

The Problem

Traditional ServiceNow development means:

• ❌ Manual navigation through complex web UIs
• ❌ Context switching between modules
• ❌ Repetitive clicking and form filling
• ❌ Limited to browser-based development
• ❌ Locked into specific AI providers

The Solution

Snow-Flow provides:

• ✅ Conversational Development: Build through natural language
• ✅ 411 ServiceNow Tools: Complete API coverage via 2 unified MCP servers
• ✅ 75+ LLM Providers: Claude, GPT-4o, Gemini, Llama, Mistral, DeepSeek, Groq, or ANY provider via OpenCode
• ✅ 100% Free Option: Ollama, LM Studio, Jan - run models locally offline
• ✅ Local Development: Pull artifacts to local files, edit with native tools
• ✅ Local ML Training: TensorFlow.js neural networks (runs locally, trains on ServiceNow data)
• ✅ Universal Compatibility: Works with OpenCode, Claude Code, Cursor, Windsurf, Continue.dev, Cline

🚀 Quick Start (3 Minutes)

Installation

# Install Snow-Flow globally
npm install -g snow-flow

# Initialize your project
snow-flow init

Snow-Flow will automatically:

• ✅ Detect and install OpenCode (recommended AI coding platform)
• ✅ Generate configuration files (.env, .opencode/config.json)
• ✅ Set up 2 MCP servers with 411 ServiceNow tools
• ✅ Create documentation (AGENTS.md, README.md)

Configuration

# Edit .env file with your credentials
vi .env

Required settings:

# ServiceNow OAuth
SNOW_INSTANCE=your-instance.service-now.com
SNOW_CLIENT_ID=your-client-id
SNOW_CLIENT_SECRET=your-client-secret

# Choose ONE LLM provider
DEFAULT_LLM_PROVIDER=anthropic  # or openai, google, ollama, etc.

# Option 1: Claude Pro/Max (leave API key empty)
ANTHROPIC_API_KEY=

# Option 2: Pay-per-use API
ANTHROPIC_API_KEY=sk-ant-your-key

# Option 3: Free local (Ollama)
DEFAULT_LLM_PROVIDER=ollama
OLLAMA_BASE_URL=http://localhost:11434
DEFAULT_OLLAMA_MODEL=llama3.3

Authentication

# One command authenticates EVERYTHING
snow-flow auth login

What happens:

1. Claude Pro/Max users: Browser opens automatically for Anthropic login
2. API key users: Skips to step 2
3. ServiceNow: Browser opens for OAuth authentication
4. Done: You're ready to develop!

OpenCode Setup

For detailed OpenCode configuration (MCP servers, environment variables, troubleshooting), see:

📖 OPENCODE-SETUP.md - Complete OpenCode configuration guide

Quick checklist:

• ✅ Use "environment" (not "env") in opencode-config.json
• ✅ Run npm run build to create dist/ directory
• ✅ Verify MCP servers load when OpenCode starts
• ✅ Test tools actually execute (not just show code snippets)

Start Developing

# Use Snow-Flow swarm for any ServiceNow task
snow-flow swarm "create incident dashboard widget with real-time charts"

# Or launch OpenCode with Snow-Flow tools pre-loaded
opencode

💡 LLM Provider Options

Snow-Flow works with 75+ LLM providers through OpenCode and Models.dev. Choose ANY model that fits your needs!

🌟 Popular Providers

🔓 Full Provider List (75+)

Via OpenCode + Models.dev:

• Anthropic: Claude Sonnet 4.5, Claude Haiku 4.5, Claude Opus 4.1
• OpenAI: GPT-5, GPT-4o, GPT-4o-mini
• Google: Gemini 2.0 Flash, Gemini 2.5 Pro, Gemini 2.5 Flash
• Meta: Llama 3.3 (8B, 70B), Llama 3.2, Llama 3.1 (405B)
• Mistral AI: Mistral Large, Mistral Medium, Mistral Small, Codestral
• Cohere: Command R, Command R+
• DeepSeek: DeepSeek Coder, DeepSeek Chat
• Groq: Ultra-fast inference for Llama, Mixtral, Gemma
• Perplexity: Web search-enabled models
• OpenRouter: 200+ models aggregator
• Local Models: Ollama, LM Studio, Jan, LocalAI, vLLM
• Enterprise: Azure OpenAI, AWS Bedrock, GCP Vertex AI
• + 60 more providers! See Models.dev for complete list

💡 Quick Setup Examples

Claude Pro/Max (No API key needed):

# .env
DEFAULT_LLM_PROVIDER=anthropic
ANTHROPIC_API_KEY=  # Leave empty!

# Authenticate
snow-flow auth login  # Opens browser automatically

OpenAI GPT-4o:

# .env
DEFAULT_LLM_PROVIDER=openai
OPENAI_API_KEY=sk-...  # From platform.openai.com

100% Free Local (Ollama):

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh
ollama pull llama3.3

# .env
DEFAULT_LLM_PROVIDER=ollama
OLLAMA_BASE_URL=http://localhost:11434

Custom/Fine-tuned Models:

# Any OpenAI-compatible endpoint works!
DEFAULT_LLM_PROVIDER=openai-compatible
OPENAI_BASE_URL=https://your-endpoint.com/v1
OPENAI_API_KEY=your-key

🎯 Recommendations by Use Case

💡 Pro Tip: Mix providers! Use cheap models for testing, expensive models for complex widgets.

🎯 ServiceNow OAuth Setup

Snow-Flow uses OAuth 2.0 for secure ServiceNow access:

1. Navigate to: System OAuth → Application Registry
2. Click: New → "Create an OAuth API endpoint for external clients"
3. Configure:
   • Name: Snow-Flow Integration
   • Redirect URL: http://localhost:3005/callback ⚠️ Must be exact!
   • Refresh Token Lifespan: 0 (unlimited)
   • Access Token Lifespan: 1800 (30 minutes)
4. Save → Copy Client ID and Client Secret
5. Add to .env (see Configuration above)

🔥 Core Features

Conversational Development

# Create widgets through conversation
snow-flow swarm "create incident dashboard with priority charts, SLA timers, and mobile-responsive design"

# Automate data operations
snow-flow swarm "create business rule to auto-assign incidents based on category and priority"

# Train ML models conversationally
snow-flow swarm "train incident classifier on last 6 months of data"

411 ServiceNow Tools (2 MCP Servers)

ServiceNow Unified Server (235+ tools):

• Complete CRUD operations on any table
• Widget deployment & validation
• UI Builder components
• Business rules & client scripts
• UI actions & policies
• REST/SOAP integrations
• Security & compliance
• Native ServiceNow ML & predictive analytics integration
• Local artifact development

Snow-Flow Orchestration (176+ tools):

• Multi-agent coordination
• Local ML training (TensorFlow.js - runs on your machine)
• Memory management
• Performance tracking
• Task orchestration

Local Development Bridge

# Pull ServiceNow artifacts to local files
snow_pull_artifact({ sys_id: 'widget_sys_id', table: 'sp_widget' })

# Edit locally with full AI coding assistant capabilities:
# - Multi-file search & replace
# - Advanced refactoring
# - Git integration
# - Native code navigation

# Push back with validation
snow_push_artifact({ sys_id: 'widget_sys_id' })

Local ML Training (Alternative to ServiceNow PI)

⚠️ Important: ML training runs locally on your machine, not in ServiceNow. This is an alternative to ServiceNow's Predictive Intelligence (PI) license.

How it works:

1. Fetch data from ServiceNow via OAuth2 API
2. Train TensorFlow.js models locally (Node.js)
3. Models saved in .snow-flow/ml-models/
4. Make predictions locally
5. Write results back to ServiceNow via API

Examples:

# Incident classification with LSTM networks (fetches ~5000 incidents)
snow-flow swarm "train incident classifier on description and short_description fields"

# Change risk assessment (fetches historical changes)
snow-flow swarm "predict change risk for upcoming releases"

# Anomaly detection (fetches system metrics)
snow-flow swarm "detect anomalies in incident patterns"

Key Differences from ServiceNow PI:

Best For:

• ✅ Development/testing without PI license
• ✅ One-off analyses and predictions
• ✅ Learning ML concepts with ServiceNow data
• ❌ Production deployments (use ServiceNow PI instead)

Native ServiceNow PI Solution Builder

🆕 Build Predictive Intelligence solutions conversationally! Create, train, and activate production-ready ML models directly in ServiceNow through natural language.

⚠️ Important: These tools build native ServiceNow PI solutions that run INSIDE ServiceNow. Requires Predictive Intelligence license.

Complete Workflow:

# 1. Create PI solution definition
snow-flow swarm "create predictive intelligence solution to predict incident category based on description"

# 2. Train the model (runs in ServiceNow)
snow-flow swarm "train the incident category predictor solution"

# 3. Monitor training progress
snow-flow swarm "check training status of incident category predictor"

# 4. Activate for production use
snow-flow swarm "activate incident category predictor solution"

# 5. Make predictions
snow-flow swarm "predict category for incident INC0010001"

Available PI Tools:

• snow_create_pi_solution - Create solution definitions (classification, regression, similarity, clustering)
• snow_train_pi_solution - Trigger training (10-30 min in ServiceNow)
• snow_monitor_pi_training - Check training progress and metrics
• snow_activate_pi_solution - Activate for production predictions
• snow_list_pi_solutions - List all PI solutions with metrics

Example Use Cases:

# Incident auto-categorization
snow-flow swarm "build PI solution to auto-categorize incidents based on short_description and description"

# Change risk prediction
snow-flow swarm "create regression model to predict change implementation duration"

# Similar incident finder
snow-flow swarm "build similarity solution to find related incidents for knowledge reuse"

# Work assignment optimization
snow-flow swarm "create PI solution to predict best assignment group for new incidents"

Key Benefits:

• ✅ Native ServiceNow: Models run directly in ServiceNow (no external dependencies)
• ✅ Production Ready: Fully integrated with ServiceNow workflows
• ✅ Auto-retrain: Automatic retraining on schedule
• ✅ High Performance: Enterprise-grade ML infrastructure
• ✅ Conversational: Build complex ML solutions through natural language

Comparison:

📋 Examples

Widget Development

snow-flow swarm "create incident dashboard with:
- Real-time counts by priority
- ML-powered trend analysis
- Interactive charts (Chart.js)
- Mobile-responsive design
- Dark mode support"

Process Automation

snow-flow swarm "create business rule for incident auto-assignment with:
- Auto-categorization based on description keywords
- Priority-based assignment rules
- Team workload balancing
- Email notifications to assignees"

Security Automation

snow-flow swarm "perform SOX compliance audit on:
- User access controls
- Change management ACLs
- Data encryption settings
Generate remediation plan for any issues"

Data Analysis

snow-flow swarm "analyze incident management process and identify:
- Bottlenecks
- Resolution time trends
- Common root causes
Provide optimization recommendations"

🤖 Multi-Agent Coordination

Snow-Flow uses OpenCode's native Task() system for intelligent agent teams:

Available Agent Types:

• widget-creator - Service Portal widget development
• security-specialist - ACL and compliance validation
• ml-specialist - Machine learning model training
• performance-optimizer - Code and query optimization
• integration-specialist - REST/SOAP integrations

Automatic coordination:

• Solo Mode: Simple tasks handled directly
• Team Mode: Complex tasks spawn specialized agent teams
• Parallel Execution: Multiple agents work simultaneously

🔧 Using with Other AI Coding Assistants

Snow-Flow works with any MCP-compatible AI coding assistant:

Quick setup:

# After snow-flow init, copy MCP config to your AI assistant
snow-flow init

# For Cursor
cp .opencode/config.json .cursor/mcp.json

# For Windsurf
cp .opencode/config.json mcp_config.json

# Restart your AI assistant → 411 Snow-Flow tools available!

📚 Advanced Configuration

Environment Variables

See .env.example for all options. Key variables:

# ServiceNow
SNOW_INSTANCE=your-instance.service-now.com
SNOW_CLIENT_ID=your-client-id
SNOW_CLIENT_SECRET=your-client-secret

# LLM Provider
DEFAULT_LLM_PROVIDER=anthropic  # or openai, google, ollama

# Provider-specific
ANTHROPIC_API_KEY=  # Leave empty for Claude Pro/Max
OPENAI_API_KEY=your-key
GOOGLE_API_KEY=your-key
OLLAMA_BASE_URL=http://localhost:11434
OLLAMA_MODEL=llama3.3

# Snow-Flow Features
LOG_LEVEL=info
ENABLE_PERFORMANCE_TRACKING=true
ENABLE_MEMORY_SYSTEM=true

Authentication Commands

# Login (handles LLM + ServiceNow)
snow-flow auth login

# Check status
snow-flow auth status

# Logout
snow-flow auth logout

Common Issues

"Could not find artifact with sys_id..."

• Cause: OAuth token expired
• Solution: snow-flow auth login

"Invalid redirect_uri"

• Cause: Redirect URL mismatch in ServiceNow
• Solution: Ensure redirect URL is exactly http://localhost:3005/callback

🎓 Use Cases

Development Teams

• Conversational development: Build through natural language
• Multi-LLM flexibility: Choose best provider per task
• Local tool integration: Use familiar development environments
• Real-time validation: Immediate feedback on artifacts

ServiceNow Administrators

• Universal operations: Query and manage any table
• Automated analysis: ML-powered insights
• Batch operations: Large-scale changes efficiently
• Cost optimization: Use local models for routine tasks

Solution Architects

• Pattern recognition: Identify and apply best practices
• Cross-module integration: Coordinate complex solutions
• Risk assessment: Analyze impact before implementation
• Provider selection: Choose optimal LLM per phase

⚙️ Requirements

• Node.js: 18.0.0 or higher
• ServiceNow: Any supported version
• LLM Provider: Any of the supported options above
• AI Coding Assistant: OpenCode (auto-installed) or others
• Memory: 4GB RAM recommended for ML training
• Storage: 1GB free space for models and artifacts

🆘 Support & Resources

• Documentation: Generated during snow-flow init
• GitHub: https://github.com/groeimetai/snow-flow
• NPM: https://www.npmjs.com/package/snow-flow
• Issues: Bug reports and feature requests welcome

🏢 Open Source vs Enterprise Features

Snow-Flow follows an Open Core business model. The core framework is 100% open source (Elastic License 2.0), with optional enterprise features available for organizations requiring advanced integrations and support.

Open Source (Elastic License 2.0) - 100% Free Forever

The current Snow-Flow release includes:

• ✅ 411 ServiceNow Tools - Complete API coverage via 2 MCP servers
• ✅ 75+ LLM Providers - Claude, GPT-4o, Gemini, Llama, Ollama, Groq, Mistral, DeepSeek, etc.
• ✅ Conversational Development - Build widgets, flows, workspaces through natural language
• ✅ Local Development Bridge - Pull/push artifacts, edit with native tools
• ✅ Local ML Training - TensorFlow.js neural networks (experimental, runs on your machine)
• ✅ Native ServiceNow ML - Predictive Intelligence solution builder (create/train/activate/predict)
• ✅ Widget Deployment & Validation - Coherence checking, ES5 validation
• ✅ UI Builder Integration - Complete Now Experience Framework development
• ✅ Multi-Agent Orchestration - SPARC methodology, swarm coordination
• ✅ Update Set Management - Full lifecycle, auto-sync current update set
• ✅ Background Script Execution - With output capture and auto-confirm mode
• ✅ Universal Compatibility - Works with OpenCode, Claude Code, Cursor, Windsurf, Continue.dev, Cline

Use for: Solo developers, small teams, learning, proof-of-concepts, community projects

Enterprise Features (Commercial License) 🔒 - Coming Soon

Premium integrations for enterprise ServiceNow teams:

• 🔒 Jira Backlog Sync - Bi-directional sync with AI-powered requirement parsing

  • Auto-import stories/epics from Jira to ServiceNow tasks
  • Real-time webhook updates (Jira ↔ ServiceNow)
  • AI extracts acceptance criteria, technical requirements
  • Developer workflow: snow-flow dev start SNOW-456 loads full Jira context

• 🔒 Azure DevOps Integration - Complete ALM integration

  • Work item sync (User Stories, Tasks, Bugs)
  • Pull request tracking to ServiceNow
  • Build pipeline status integration
  • Test results auto-documented

• 🔒 Confluence Documentation Sync - Auto-sync technical documentation

  • Confluence pages → ServiceNow knowledge articles
  • Technical specs → Implementation guides
  • Architecture diagrams → Attachment sync

• 🔒 Enterprise SSO/SAML - Single sign-on integration

  • SAML 2.0, OAuth 2.0, OIDC
  • Active Directory/LDAP integration
  • Role-based access control (RBAC)

• 🔒 Advanced Audit Logging - Comprehensive activity tracking

  • Every AI action logged with audit trail
  • Compliance reporting (SOX, GDPR, HIPAA)
  • Tamper-proof log storage

• 🔒 SLA-Backed Support - Enterprise support agreements

  • 24/7 support with 4-hour response SLA
  • Dedicated technical account manager
  • Priority feature requests
  • Custom integration development

• 🔒 Advanced ML Features - Enhanced predictive intelligence

  • Auto-retrain on schedule
  • A/B testing for ML models
  • Multi-model ensembles
  • Custom feature engineering

Use for: Enterprise teams (20+ developers), consulting firms (Capgemini/Accenture/Deloitte), ISVs, production deployments

Pricing (Coming Q2 2025)

Enterprise features will be available via commercial license:

• Professional: €499/month (5 developers)
• Team: €999/month (20 developers)
• Enterprise: €1,999/month (unlimited developers)

All tiers include:

• All open source features
• Enterprise integrations (Jira/Azure DevOps/Confluence)
• Enterprise SSO/SAML
• Audit logging
• Email support (SLA varies by tier)

Early Access: Contact sales@snow-flow.dev for beta access and pilot programs.

Strategic Partners: Consulting firms and ISVs - contact us for partnership opportunities.

Why Open Core?

1. Community-Driven Innovation - Core features benefit from community contributions
2. Transparent Development - All core development happens in public
3. No Lock-In - Open source core means you can self-host and modify
4. Sustainable Development - Enterprise revenue funds ongoing development
5. Enterprise Confidence - Commercial license provides support and guarantees

Roadmap

Q1 2025 (Current Release - Open Source)

• ✅ Native ServiceNow Predictive Intelligence tools
• ✅ UI Builder complete integration
• ✅ 75+ LLM provider support
• ✅ Local development bridge

Q2 2025 (Enterprise Launch)

• 🔒 Jira Integration (beta)
• 🔒 Azure DevOps Integration (beta)
• 🔒 Enterprise SSO/SAML
• 🔒 Audit logging

Q3-Q4 2025 (Enterprise Expansion)

• 🔒 Confluence Integration
• 🔒 Advanced ML features
• 🔒 Multi-tenant architecture
• 🔒 White-label options

2026+

• More enterprise integrations (GitHub, GitLab, Linear, Asana)
• Advanced compliance features
• AI-powered code review
• Automated testing frameworks

See MONETIZATION_STRATEGY.md for complete business model details.

🏢 Snow-Flow Enterprise

Enterprise-grade features for large-scale ServiceNow development

Available Now in Beta

Snow-Flow Enterprise adds advanced capabilities for enterprise teams:

🔗 Jira Integration

• Bidirectional Sync: Keep Jira and ServiceNow in perfect sync
• Automated Workflows: Trigger actions across both platforms
• Field Mapping: Customize data transformation
• Comment Sync: Unified communication across teams
• Attachment Support: Share files seamlessly

🤖 Advanced ML Models

• Deep Learning: LSTM, CNN, transformer-based models
• Auto-ML: Automatic model selection and tuning
• Model Versioning: Track and manage model lifecycles
• Explainable AI: SHAP/LIME analysis for predictions
• A/B Testing: Compare models in production

🎯 Priority Support

• 4-24 hour response times based on tier
• Dedicated account manager (Enterprise tier)
• On-site training available
• Custom integration development
• SLA guarantees (99.9% uptime)

License Tiers

Quick Start

# Install enterprise package
npm install -g @snow-flow/enterprise

# Configure license
export SNOW_LICENSE_KEY="SNOW-ENT-XXXXX"

# Verify installation
snow-flow license status

# Start using enterprise features
snow-flow swarm "Sync Jira project PROJ to ServiceNow"

Example: Jira Sync

# Sync entire backlog
snow-flow swarm "Sync all issues from PROJ project in Jira to ServiceNow"

# Search and sync
snow-flow swarm "Find all high-priority bugs in Jira and create ServiceNow incidents"

# Add comments
snow-flow swarm "Add comment to Jira PROJ-123: Resolved in ServiceNow INC0010001"

Documentation

• Enterprise Documentation - Complete guide
• Quick Start Guide - Get started in 5 minutes
• License Agreement - Commercial terms
• MCP Tools Reference - Enterprise tools

Get Enterprise Access

30-Day Free Trial available for all tiers!

• 📧 Email: sales@snow-flow.dev
• 💬 Schedule Demo: calendly.com/snow-flow
• 📞 Call: +1 (555) 123-4567

📄 License

Core Framework: Elastic License 2.0 - see LICENSE file for details.

What this means:

• ✅ Free to use, modify, and redistribute
• ✅ Can be used commercially
• ✅ Full source code access
• ⚠️ Cannot provide Snow-Flow as a managed service to third parties
• ⚠️ Cannot remove or obscure licensing/copyright notices

Trademark: "Snow-Flow" name and logo are protected trademarks - see TRADEMARK.md.

Enterprise Features: Commercial license required - contact sales@snow-flow.dev.

Snow-Flow: Conversational ServiceNow Development • 411 Tools • 2 MCP Servers • 75+ LLM Providers • Universal Compatibility

Get started in 3 minutes:

npm install -g snow-flow
snow-flow init
snow-flow auth login
snow-flow swarm "create incident dashboard"