statistics

Deterministic statistical computation for AI agents.

Agents asked to compute a mean, median, or standard deviation will often approximate or round incorrectly. On small datasets the errors are subtle. On large ones they compound. Euclid's statistics tool computes exact results on real data, so your agent's summaries and reports are verifiably correct.

Live

euclid / statistics

statistics("mean", [84200, 91500, 78300, 96100]){ "result": "87525" }

statistics("percentile", [120, 340, 200, 150, 180, 95, 220, 310, 170, 190], 90){ "result": "313" }

statistics("median", [72, 85, 91, 68, 77, 95, 82, 88]){ "result": "83.5" }

statistics("std", [10.02, 10.05, 9.98, 10.01, 9.97, 10.03]){ "result": "0.030331..." }

Operations

meanmedianmodestandard deviationvarianceminmaxsumpercentile

Examples

What it looks like in practice

Quarterly sales performance review

statistics("mean", [84200, 91500, 78300, 96100])

{ "result": "87525" }

Identifying the 90th percentile response time

statistics("percentile", [120, 340, 200, 150, 180, 95, 220, 310, 170, 190], 90)

{ "result": "313" }

Student grade distribution

statistics("median", [72, 85, 91, 68, 77, 95, 82, 88])

{ "result": "83.5" }

Manufacturing quality control

statistics("std", [10.02, 10.05, 9.98, 10.01, 9.97, 10.03])

{ "result": "0.030331..." }

Real estate price analysis

statistics("mode", [450000, 520000, 450000, 480000, 450000, 510000])

{ "result": "450000" }

Weekly calorie tracking

statistics("sum", [2100, 1850, 2300, 1950, 2050, 2400, 1900])

{ "result": "14550" }

Who uses this

Built for agents across every domain

fn()

Product manager

An AI agent summarising A/B test results where the difference between 2.3% and 2.8% conversion determines whether a feature ships.

fn()

HR analyst

Computing salary percentiles across departments to identify pay equity gaps in a compensation review.

fn()

Sports analyst

An agent computing player performance statistics where a wrong standard deviation misranks draft picks.

fn()

Quality engineer

Monitoring manufacturing tolerances. A wrong variance calculation means defective parts reach customers.

fn()

Researcher

An AI assistant verifying descriptive statistics before a paper submission. Wrong numbers retract publications.

fn()

Personal trainer

An AI fitness agent tracking client progress over time where accurate averages and trends matter for programming adjustments.

Try statistics on Euclid Cloud

One MCP connection. Every computation your agents need.

Get started