Telemetry

Project Overview

Telemetry is Jetpack Labs’ internal reference platform designed to build, validate, and refine architectural patterns for multi-bot AI orchestration systems. Rather than abstractly designing multi-agent systems based on theory, Jetpack Labs chose to dogfood the architecture by building a real, production-grade platform that handles complex, multi-step workflows with strict guardrails, observability, and tenant isolation.

The platform powers specialized AI agents (Article Writer, Social Media Manager, SEO Analyst, PR Specialist, Sales Agent, Script Agent, and more) that collaborate to generate high-quality content at scale. It manages real-time async processing pipelines where multiple agents work on the same task, handle intelligent routing based on user intent, and integrate knowledge bases via RAG for contextual awareness.

Key Features

1. Multi-Agent Orchestration Engine

   • AgentInterface contract for type safety
   • AbstractAgent base with shared utilities
   • MasterAgent router for intelligent dispatching
   • Inline prompt building (no config templates)
   • Strict JSON output validation
   • 10+ specialized bot types with minimal boilerplate

2. Real-Time Frontend & WebSocket

   • Laravel Reverb for secure WebSocket (WSS)
   • Livewire Volt/Flux components
   • Real-time conversation updates and progress tracking
   • Streaming agent responses
   • 99.9% uptime with graceful degradation

3. Async Job Pipeline & Processing

   • Database-driven queue for reliability
   • Chainable job pipelines
   • Content generation workflow: Outline → Sections → SEO → Finalize
   • Image generation (hero + social)
   • Intermediate artifact storage
   • Automatic retry with exponential backoff

4. AI Safety & Guardrails

   • Role-based action constraints
   • Strict prompt engineering (low temperature)
   • JSON schema validation
   • Deterministic model settings
   • Per-bot ability whitelisting
   • Workspace-level guardrail overrides

5. Knowledge Base & RAG

   • Per-workspace knowledge base context
   • Semantic search and retrieval
   • Conversation history as context
   • Workspace guidelines injection
   • Web context for recent data

6. Multi-Tenant Architecture

   • Workspace-scoped queries & isolation
   • Per-workspace bot customization
   • Conversation-scoped guidelines
   • Model response logging for audit
   • Structured event telemetry
   • Performance & cost tracking

Technologies Used

• Backend: Laravel, Laravel Reverb (WebSocket), Claude API, OpenAI (image generation), MySQL, Redis, Docker
• Frontend: Livewire Volt/Flux, TypeScript, Vite
• Testing: PHPUnit
• Architecture: Multi-tenant SaaS, Real-time async job queues, Extensible agent framework, Role-based guardrails, Knowledge base & RAG

Key Results

• 4× Faster Content Generation Workflows - Parallelized agent pipelines generate multi-step content in a fraction of the time
• 90% Code Reuse via Extensible Architecture - New bot types can be added with minimal boilerplate
• 10+ Specialized AI Agents - Each with deterministic prompts, role-based action constraints, and strict JSON validation
• Real-Time UI with 99.9% Uptime - Sub-100ms WebSocket updates with thousands of concurrent connections
• Reference Architecture for Client Projects - Telemetry patterns directly inform client engagements across manufacturing, supply chain, and education platforms

Technical Insights

1. Inline Prompt Building Beats Configuration Templates - PHP heredoc strings for prompts proved more maintainable than config-driven templates
2. Strict JSON Output Schemas Eliminate Parsing Complexity - Reduced parsing errors by 99%
3. WebSocket Requires Graceful Degradation - Automatic reconnection, request queuing, and polling fallbacks
4. Async Pipelines Need Idempotent Job Steps - Every job step must be idempotent and track its own completion state
5. Guardrails Are Cheaper Than Error Recovery - Prevents out-of-scope agent behavior before reaching the LLM

Impact on Jetpack Labs’ Client Work

Telemetry’s architectural patterns directly benefit client engagements:

• Manufacturing Quoting Engine - Routes RFQs to specialized agents, reducing quote turnaround from days to hours
• Supply Chain Optimization - Coordinates multi-step planning with real-time stakeholder updates
• Education Student Success Platform - Provides personalized AI tutors with strict compliance and privacy boundaries
• Logistics Route Planning - Integrates external data for AI-driven routing, resulting in 15% cost savings per route
• Sales & Marketing Automation - Dispatches specialized agents for lead qualification and follow-up, increasing conversion by 20%