Live on ChatGPT and Claude, WhenPeak now powers AI agent scheduling

Your AI agent doesn't know when you peak. Now it can.

WhenPeak is the performance intelligence layer for AI agents and the humans they serve. One API call tells your agent when to schedule deep work, protect the dip, and prepare for the moments that matter.

Live on ChatGPTLive on ClaudeOpenAPITwo-Process Model222 unit tests
Today's performance curve
Third Bird · Good sleep · Woke 7am
Peak 10:00am
Dip 2:00pm
Peak 2 4:00pm
15–30%Faster on complex tasks
20–40%Fewer errors at peak
2–5 yrsFaster to director level

Predictions update daily. Connect Apple Health for your personal curve.

CAREER IMPACT CALCULATOR
ILLUSTRATIVE ONLY

Your biology is compounding.
Are you timing it right?

Set your salary, expected promotion timeline, and industry. See how cognitive protocol execution diverges from the average trajectory over five years.

Protocol 5-yr total
$1,019,385
Average 5-yr total
$888,416
Earnings advantage
+$130,969
Protocol: time to promotion
2.4 yrs
Average: time to promotion
3.0 yrs
$50K$500K
2.5 yrs5 yrs

Tech / Startup model. Protocol: 7% annual raises + 22% promo bump, promotion 20% earlier. Average: 4% annual raises + 15% promo bump. Sources: McKinsey high-performance research; organisational psychology meta-analyses on merit allocation and promotion velocity. For illustrative purposes only. Not financial advice. Projections are hypothetical. Actual outcomes depend on employer, role, performance, and economic conditions. Individual results vary significantly.

One platform. Two ways to experience it.

Whether you're building AI agents or optimising your own performance, WhenPeak provides the same prediction engine, via API or directly.

For developers

Give your agent a cognitive layer

AI agents schedule meetings, prioritise tasks, and draft responses, but they don't know if their human is cognitively ready for what's being scheduled. A single WhenPeak API call gives your agent that context: peak window, dip time, daily readiness score, and chronotype.

  • Agent calls /api/v1/predict at the start of the user's day
  • Receives peak window, dip time, DPS, and upgrade prompts
  • Schedules work, meetings, and focus blocks accordingly
POST /api/v1/predict
→ peak_1: "10:00", dip: "14:00",
  peak_2: "17:00", dps: 74
For individuals

Know your cognitive state before your day begins

Connect Apple Health or answer three questions about last night's sleep. WhenPeak returns your 24-hour performance curve: when to do your most important work, when to rest, and how last night's sleep changed today's trajectory.

  • Daily performance curve from your actual sleep data
  • Peak, dip, and secondary window for every day
  • Works in ChatGPT, Claude, or the mobile app
Today's performance curve
Third Bird · Good sleep · Woke 7am
Peak 10:00am
Dip 2:00pm
Peak 2 4:00pm
Try in ChatGPT →Mobile app coming
Developer Platform

Build agents that understand their humans

Your agent knows the user's calendar, email, and tasks. It doesn't know if they slept well or when their brain peaks. WhenPeak is the missing context layer, one API call, clean JSON, works with any framework or LLM.

No auth required for first callOpenAPI spec availableWorks with GPT Actions
curl -X POST https://api.whenpeak.com/api/v1/predict \
  -H "Content-Type: application/json" \
  -d '{
    "sleep_time": "23:30",
    "wake_time": "07:00",
    "sleep_quality": "good",
    "exercise_yesterday": true,
    "exercise_timing": "morning"
  }'
PREDICT

Per-user performance curves

Daily peak, dip, secondary window, DPS, and chronotype. Stateless (no auth) or personalised (authenticated, improves over 30 days).

METER

Developer dashboard

See your end-users' aggregated performance metrics, API usage, and data confidence levels, so you know how well the model is calibrated for your user base.

SCALE

Multi-user architecture

Provision isolated end-user accounts under your developer key. Each user gets their own sleep history, chronotype, and personalised curve. Growth and Scale tiers.

Works where you already work

WhenPeak runs inside the tools your users are already in.

ChatGPT

ChatGPT

Available now as a GPT. Users ask in plain English, WhenPeak predicts their peak. No app required.

Live
Claude

Claude

MCP skill deployed. Claude can call WhenPeak directly to answer performance questions for its users.

Live
</>

Any AI agent

OpenAPI spec published. Any LLM with function calling can integrate WhenPeak in under an hour.

Open APILive
Mobile

Mobile app

Direct access to your performance curve, sleep history, and 7-day outlook. Connect Apple Health or WHOOP.

Coming soon
RESEARCH BASIS

Built on peer-reviewed science

The Two-Process Model (Borbély 1982) governs every prediction. Process C tracks your circadian clock. Process S tracks homeostatic sleep pressure. Their interaction produces your 24-hour performance curve, calibrated to your chronotype.

WhenPeak's model is built on peer-reviewed chronobiology and a review of performance research from leading institutions.

  • -Sleep quality, duration, consistency, and debt
  • -HRV relative to your 30-day personal baseline
  • -Exercise timing and chronotype classification
PerformancePeak 10:00Dip 14:00Recovery 16:3006:0010:0014:0018:0022:00
HarvardStanfordUC BerkeleyOxfordColumbiaCambridgeMITNYUNIH

Simple pricing

Start free. Scale when you're ready.

Free
$0/ month

For developers and early exploration

  • 1,000 API calls / month
  • Stateless predictions
  • OpenAPI spec
  • Community support
Most popular
Growth
$49/ month

For products with real users

  • 50,000 API calls / month
  • Authenticated user predictions
  • 30-day personal baselines
  • 7-day performance forecasts
  • Webhook support
  • Email support
Scale
$199/ month

For high-volume integrations

  • 500,000 API calls / month
  • Everything in Growth
  • White-label available
  • Dedicated support
  • SLA

Enterprise or white-label? hello@whenpeak.com

Common questions

Answers to what people ask before they integrate.

Your peak is predictable.

Start with one API call. No account required.