PROTOCOL ENGINE v1.0

Messaging protocols for health interventions

Design branching message sequences, schedule timezone-aware delivery, and monitor queue health — built for research teams running intervention studies at scale.

Go to Dashboard Learn More

protocol_v11.yaml

# V11 Quit-Smoking Protocol
protocol:
  name: "QuitTxt 2025"
  nodes: 1,024
  templates: 2,048 (en/es)

flow:
  entry → daily_checkin
  daily_checkin → on_track | response == "good"
  daily_checkin → slip_support | response == "slip"
  daily_checkin → helpnow | response == "help"
  daily_checkin → timeout_nudge | no_response 24h

delivery:
  channels: [sms, push]
  window: 08:00–20:00 participant_tz
  retry: exponential(max=8)

About Cadence

A complete platform for automated health messaging

Cadence is a messaging protocol management system designed for research teams running behavioral health intervention studies. It handles the full lifecycle — from designing branching message protocols to delivering SMS and push notifications on personalized schedules.

Built on the V11 protocol specification used in quit-smoking interventions, Cadence supports complex conditional flows, bilingual templates, timezone-aware scheduling, and real-time monitoring of message delivery across thousands of participants.

Visual Protocol Editor

Drag-and-drop graph editor for building branching message flows with conditional edges and variable-based routing.

Participant-Centric Delivery

Each participant gets messages in their timezone, preferred language, and on their chosen channel — with automatic fallback.

Research-Grade Security

HIPAA-conscious design with AES-256 encryption, role-based access, and full audit trails on every interaction.

Message Lifecycle

From protocol definition to delivery

Define

Build node graph with message templates, conditions, and branching paths

Enroll

Add participants with timezone, language preference, and channel settings

Schedule

Engine evaluates protocol state and queues messages within delivery windows

Deliver

Messages sent via SMS/push with retry logic, receipts, and failure handling

Capabilities

Technical specifications

Protocol Engine

V11 branching protocol with conditional paths

Node types: message, check-in, escalation, checkout
Edge conditions: timeout, response, variable match
Session state persisted in Redis (sub-ms lookup)
Bilingual templates with variable interpolation

Scheduler

Timezone-aware delivery with retry logic

Configurable delivery windows per participant
Exponential backoff: up to 8 retries
Batch processing: 100 messages per cycle
Oldest-pending monitoring with auto-requeue

Delivery

Multi-channel with automatic fallback

SMS via Twilio (primary channel)
Push notifications via Firebase FCM
Channel fallback on delivery failure
Per-message delivery receipts and status tracking

Participants

Cohort management with full audit trails

Bulk enrollment via API or admin dashboard
Protocol stage tracking per participant
HELPNOW escalation and SLIP detection
Withdrawal and suspension workflows

Security

Research-grade data protection

JWT auth with HttpOnly secure cookies
API key isolation (timing-safe comparison)
AES-256 PII encryption at rest
Role-based access: admin, researcher, observer

Analytics

Real-time queue health and delivery metrics

Sent / pending / failed / in-progress counters
Per-project and global scheduler dashboards
Message latency tracking (oldest pending)
Exportable delivery logs for analysis

Architecture

System overview

CLIENT

Next.js 14 (App Router)React Flow graph editorZustand state

API

FastAPI + async/await~70 REST endpointsSlowAPI rate limiting

DATA

PostgreSQL (19 models)Redis (sessions + cache)Alembic migrations

DELIVERY

Background workerTwilio SMSFirebase FCM push

REST API — Key Endpoints

METHOD	ENDPOINT
POST	`/v1/auth/login` Authenticate and set cookie
GET	`/v1/projects` List all projects
POST	`/v1/projects/:id/nodes` Create protocol node
POST	`/v1/protocol/session/start` Start participant session
POST	`/v1/protocol/session/respond` Process participant response
GET	`/v1/scheduler/health` Queue health metrics

~70 endpoints total · OpenAPI spec at /docs

Database models

SQLAlchemy ORM

~70

API endpoints

FastAPI async

1,024

Protocol nodes

V11 full import

2,048

Message templates

en/es bilingual

Ready to manage your messaging protocols?

Go to Dashboard