Notiqo CLI

Notiqo CLI scans source code repositories and generates a complete analytics tracking plan: it detects events already instrumented in your code, flags naming and coverage issues, suggests missing events with meaningful properties, and maps conversion funnels — all without requiring any manual configuration.

Features

• Framework detection: Flutter, React, Next.js, Node.js, Angular, and more via tree-sitter
• Universal language support: Dart, JavaScript, TypeScript, Python, Java, Go, Rust
• App type detection: e-commerce, SaaS, fintech, edtech, social, marketplace, content
• Business Context Collection: Interactive questionnaire captures problem statement, target users, key actions, and business objectives
• Existing event detection: reads live analytics SDK calls (Mixpanel, Amplitude, Segment, Firebase, PostHog, custom wrappers)
• Event quality analysis: flags camelCase names, missing industry-standard properties, _triggered suffix anti-patterns
• New event suggestions: surfaces gaps between what you track and what you should track
• AI-powered LLM Mode (optional --llm): LangChain agent orchestration with specialized tools
  • TechScanner: Detects technologies in codebase
  • EventAnalyzer: Detects existing events + suggests new ones from business context
  • FunnelGenerator: Generates conversion funnels based on app type
  • QualityAdvisor: Generates business insights and strategic recommendations
• Conversion funnel detection: Multiple industry templates (e-commerce, SaaS, fintech, etc.)
• Output formats: tracking-plan.md and/or tracking-plan.json
• Notiqo web sync: push the generated plan to your Notiqo dashboard

Usage

npm install
npm run build

Basic Scan (Rule-Based)

# Basic scan — generates tracking-plan.md in the target directory
node dist/cli/index.js scan .

# Output both markdown and JSON
node dist/cli/index.js scan . --format both

# Force app type (skips auto-detection)
node dist/cli/index.js scan . --app-type ecommerce

LLM-Enhanced Scan (AI-Powered)

# Enable AI enhancement with MiniMax
node dist/cli/index.js scan . --llm --minimax-key <your-api-key>

# Full pipeline with all options
node dist/cli/index.js scan . --format both --llm --minimax-key <key> --push --token <notiqo-token>

Environment Variables for Business Context

When using --llm, you can pre-fill the business context questionnaire via environment variables:

NOTIQO_PROBLEM="Fast food restaurant app for ordering" \
NOTIQO_USERS="Customers ordering food" \
NOTIQO_ACTIONS="Browse menu, add to cart, checkout, payment, track order" \
NOTIQO_OBJECTIVES="Increase sales, improve customer experience" \
NOTIQO_FUNNELS="Purchase flow, Cart recovery" \
node dist/cli/index.js scan . --llm --minimax-key <key>

| Environment Variable | Description | Required | |--------------------|-------------|----------| | NOTIQO_PROBLEM | What problem does your app solve? | Yes | | NOTIQO_USERS | Who are your users? | Yes | | NOTIQO_ACTIONS | Main actions users do (comma-separated) | Yes | | NOTIQO_OBJECTIVES | Business objectives (comma-separated) | Yes | | NOTIQO_FUNNELS | Known funnels (optional) | No |

CLI Options

| Option | Description | Default | |--------|-------------|---------| | --format <md\|json\|both> | Output file format | md | | --app-type <type> | Force app type: ecommerce, saas, fintech, edtech, social, marketplace, content | auto-detect | | --llm | Enable LLM agent orchestration | false | | --minimax-key <key> | MiniMax API key (or MINIMAX_API_KEY env var) | — | | --push | Push plan to Notiqo web | false | | --token <token> | Notiqo API token (or NOTIQO_TOKEN env var) | — |

Architecture

LLM Mode: Agent Orchestration

The LLM mode uses LangChain with a ReAct agent that orchestrates specialized tools:

┌─────────────────────────────────────────────────────────┐
│                  AnalyticsOrchestrator                    │
│  - Receives business context from user                   │
│  - Orchestrates 4 specialized tools                       │
│  - Synthesizes final tracking plan                       │
└─────────────────────────────────────────────────────────┘
         │              │              │           │
    ┌────▼────┐   ┌────▼────┐   ┌────▼────┐   ┌────▼────┐
    │  Tech   │   │  Event  │   │ Funnel  │   │ Quality │
    │ Scanner │   │ Analyzer│   │Generator│   │ Advisor │
    └─────────┘   └─────────┘   └─────────┘   └─────────┘

Tools

1. TechScanner: Scans project for config files and file extensions to detect technologies
2. EventAnalyzer: Scans codebase for existing SDK calls + generates suggestions from business context
3. FunnelGenerator: Generates conversion funnels using industry templates (ecommerce, saas, etc.)
4. QualityAdvisor: Generates business insights and strategic recommendations

Rule-Based Mode: Pipeline

1. Framework detection        (Flutter / React / Next.js / Node / Angular)
2. Code scanning              (AST-first, regex fallback)
   ├── Functions with business-relevant names
   ├── Existing analytics SDK calls
   ├── Screens / pages, routes, API calls
3. AppType detection          (ecommerce / saas / fintech / …)
4. Event generation
   ├── Existing events  → quality analysis
   └── New suggestions  → from functions, routes
5. Funnel detection
   ├── AppType-matched templates
   └── Steps resolved against events
6. Plan generation            tracking-plan.md / tracking-plan.json

Project Structure

/src
  /api
    notiqo-client.ts               # HTTP client for Notiqo web push
  /cli
    index.ts                       # Command registration (commander)
    /commands
      scan.ts                      # Scan command handler
    business-context-questionnaire.ts  # Interactive questionnaire for LLM mode
  /core
    business-context.ts            # BusinessContext types for agent orchestration
    shared-memory.ts                # Shared memory store for agent communication
    types.ts                       # Domain entities
    interfaces.ts                  # Ports: FrameworkDetector, CodeScanner, LlmClient…
    scan-orchestrator.ts           # Rule-based pipeline orchestrator
  /event-generator
    ai-function-classifier.ts      # LLM-based function classification
    rule-based-event-generator.ts  # Rule-based event generation
  /framework-detectors
    flutter-detector.ts
    react-detector.ts
    next-detector.ts
    node-backend-detector.ts
    framework-detector-resolver.ts
  /funnel-detector
    funnel-template-library.ts     # Industry templates
    funnel-generator.ts            # Funnel generation
  /llm
    langchain-client.ts           # LLM client abstraction
    orchestrator.ts                # AnalyticsOrchestrator (main agent)
    /tools
      tech-scanner-tool.ts        # TechScanner tool
      event-analyzer-tool.ts       # EventAnalyzer tool
      funnel-generator-tool.ts      # FunnelGenerator tool
      quality-advisor-tool.ts       # QualityAdvisor tool
    base-tool.ts                   # Base tool interface
    minimax-llm-client.ts        # MiniMax LLM client
    noop-llm-client.ts            # No-op fallback
  /scanner
    analytics-call-extractor.ts    # Detects SDK calls from AST
    ast-extractor.ts               # Tree-sitter traversal
    repository-scanner.ts          # File traversal and aggregation
    tree-sitter-factory.ts         # Parser factory
    language.ts                   # File extension → language mapping
  /tracking-plan
    tracking-plan-generator.ts     # Builds TrackingPlan, renders markdown/JSON
    tracking-plan-writer.ts        # Writes output files
  /utils
    question.ts                   # CLI questionnaire utility

LLM Integration

LLM enhancement is off by default. Enable with --llm.

Provider: MiniMax (https://api.minimax.io/v1/text/chatcompletion_v2) Model: MiniMax-Text-01 Key: --minimax-key <key> or MINIMAX_API_KEY env var

Agent Tools

| Tool | Purpose | Input | |------|---------|-------| | tech_scanner | Detects technologies in project | { projectPath: string } | | event_analyzer | Detects existing events + suggests new | { projectPath?: string } | | funnel_generator | Generates conversion funnels | { appType?: string } | | quality_advisor | Generates insights + recommendations | {} |

Graceful Degradation

If the LLM call fails or no key is provided, the system falls back to rule-based logic transparently.

Existing Event Detection

The scanner reads analytics SDK calls directly from source code:

| SDK | Detected patterns | |-----|-------------------| | Firebase | .logEvent(name: 'event'), FirebaseAnalytics.instance.logEvent(...) | | Mixpanel | mixpanel.track('event', {...}) | | Amplitude | amplitude.track('event', {...}) | | Segment | analytics.track('event', {...}) | | PostHog | posthog.capture('event', {...}) | | Custom | analytics.trackEvent(...), trackEvent(...), sendEvent(...) |

Output Structure

The generated tracking-plan.md has four sections:

## Existing Tracked Events
Events already instrumented in the codebase.

## Event Improvements
Naming issues and missing recommended properties.

## Suggested New Events
Gaps in coverage with suggested properties.

## Conversion Funnels
Industry-matched funnels with steps resolved to event names.

Example Output

LLM Mode

Notiqo Scan Results (LLM Mode)
==============================
Problem: Fast food restaurant app for ordering
Target users: Customers ordering food
Business objectives: Increase sales, improve customer experience

Detected technologies: 4
  - JavaScript (frontend)
  - HTML (frontend)
  - Dart (mobile)

Existing tracked events: 8
New events suggested: 5
Suggested funnels: 2
Business insights: 6
Strategic recommendations: 2

New events suggested:
- browsed
- item_added
- checkout_started
- payment_completed
- order_placed

Business insights:
- Purchase flow events are covered - good for revenue analytics
- Customer experience objective detected - consider tracking engagement events
- ...

Strategic recommendations:
- [medium] Track cart behavior
- [low] Consider NPS/survey tracking

Rule-Based Mode

Notiqo Scan Results
===================
Detected framework:    flutter
Detected app type:     ecommerce
Detected screens:      278
Detected routes:       0
Existing tracked events: 62
New events suggested:    12
Improvements found:       6
Suggested funnels:        4
Output files:
  - /project/tracking-plan.md
  - /project/tracking-plan.json