Now in Early Access

Stop guessing how much your team can handle

Clarinet measures cognitive load in real-time, prevents burnout before it happens, and optimizes task sequencing using science — not intuition.

Request Early Access See How It Works
Watch 60-second product walkthrough
49
Task Types Supported
r=0.81
Heuristic Accuracy
213
Benchmark Tasks Validated
20+
Industry Sectors
The Problem

Knowledge work is making people sick

68% of workers feel overwhelmed by work volume. Managers have no data on cognitive overload until it's too late.

Invisible Overload

Cognitive load is the #1 predictor of errors and burnout, yet no tool measures it. Managers rely on gut feeling to distribute work.

$322B Lost Annually

Workplace burnout costs the global economy $322 billion per year in turnover, absenteeism, and reduced productivity (WHO/Gallup, 2022).

One-Size-Fits-All

Neurodivergent workers (15-20% of workforce) face different cognitive challenges. Current tools ignore individual differences entirely.

How It Works

From task description to cognitive insight in seconds

Clarinet analyzes tasks through the lens of Cognitive Load Theory, the gold standard framework from educational psychology.

1

Describe Your Tasks

Enter task titles and descriptions — just like you would in Jira, Asana, or any project tool. Set your fatigue level and cognitive profile.

2

AI Evaluates Load

Our dual evaluation engine scores each task on intrinsic load, extraneous load, and context switch cost. Neurodivergence-aware, with <50ms response time.

3

Optimize & Monitor

Get an optimized task sequence, burnout risk alerts, and scheduling recommendations backed by peer-reviewed research.

The Platform

API-first cognitive load intelligence

The Clarinet API is the core product — a RESTful evaluation engine that scores any task's cognitive load in milliseconds. Built on top, our user-facing apps serve individuals and managers.

Cognitive Load API

RESTful API delivering cognitive load scores in <50ms per task. Integrate into any workflow: Jira, Asana, Slack, or your own tools. Heuristic engine with r=0.81 correlation to expert ratings.

Dual Evaluation Engine

Context-aware heuristic engine runs instantly at near-zero cost. Optional LLM-based engine provides deeper semantic analysis with delta comparison. Smart routing optimizes cost vs accuracy.

Neurodivergence Support

ADHD, ASD, AuDHD, Dyslexia, and more. Adjusted thresholds, tailored recommendations, and condition-specific scheduling strategies.

Burnout Risk Algorithm

4-factor model (wear trend, sustained load, fatigue pattern, state signals) with neurodivergence-aware thresholds. Risk levels: Low to Critical.

Individual Task Planner

Built on the API. Workers load their tasks, get cognitive load scores, optimal sequencing, and personalized advice based on their cognitive profile.

Admin Dashboard & Backoffice

For managers, team leads, and HR. Team-wide cognitive health monitoring, burnout risk scores, wear trends, and scheduling recommendations per employee.

Science & Mission

Built on cognitive science, driven by a mission

Clarinet is grounded in Cognitive Load Theory — 30+ years of peer-reviewed research in educational and occupational psychology. We don't guess; we measure.

Scientific Foundation

  • Based on Cognitive Load Theory (Sweller, Ayres & Kalyuga, 2011)
  • Validated against 213 expert-evaluated tasks across 49 types
  • Neurodivergence research: Barkley (2015), Young et al. (2014)
  • Task taxonomy: O*NET, ISCO-08, Eurofound (2022)
  • Every recommendation backed by peer-reviewed evidence

Our Mission

  • Cognitive overload is invisible, unmeasured, and costs $322B/year.
  • We're building the measurement layer that workplaces are missing.
  • A multidisciplinary team of engineers and cognitive psychologists, based in Berlin, Germany.
  • Inclusive by design: 15-20% of workers are neurodivergent — our platform adapts to them, not the other way around.

Engine Benchmark Results (v1.2)

Validated against 213 expert-evaluated tasks across 49 types and 20+ industries

Mean Absolute Error13.2 / 100
Pearson Correlationr = 0.809
Load Band Accuracy69%
LOW Band Accuracy76%
MEDIUM Band Accuracy61%
HIGH Band Accuracy71%
Tasks within ±25 pts86%
Extraneous Load Spread30.5 pts (poor vs excellent)
Pricing

Usage-based pricing that scales with value

Hybrid model: platform fee + consumption. No per-seat traps. Unlimited users on every plan.

Starter
Free
For individual contributors
  • Heuristic engine — unlimited evaluations
  • Up to 20 tasks per plan
  • Cognitive profile + ND support
  • Task sequencing
  • Community support
Get Started
Growth
€99/mo platform
2,000 LLM evaluations included • €0.05/eval overage
  • Unlimited users (no per-seat fees)
  • Admin Dashboard + burnout monitoring
  • Dual engine (heuristic + LLM)
  • Smart model routing (cost-optimized)
  • Team cognitive health & 30-day trends
  • Scheduling recommendations
  • Email support (48h SLA)
Request Access
Enterprise
Custom
Volume discounts + auto-tuned engine + committed spend
  • Everything in Growth
  • Auto-tuned engine (85% COGS reduction)
  • SSO / SAML + data residency (EU/US)
  • API access & SDK
  • Dedicated CSM + 4h SLA
  • Custom integrations (Jira, Slack, Teams)
  • On-premise deployment option
Contact Sales

Unit economics: why this model works

Our heuristic engine runs at near-zero marginal cost. LLM evaluations use smart model routing (85% lightweight models, 15% flagship for complex tasks) with prompt caching, keeping COGS at ~€0.005/eval. At €0.05/eval retail, gross margin on LLM usage exceeds 90%. The auto-tuned engine (roadmap) replaces 90% of LLM calls with local inference at €0.001/eval, expanding margins to 97%+.

<€0.001
Heuristic COGS
€0.005
LLM COGS (routed)
€0.05
LLM retail price
90%+
Gross margin

Ready to see Clarinet in action?

We're onboarding early design partners. Limited spots available.

Request Demo