Rametric Labs

Labs

Small working examplesof the mechanics behind AI systems.

The labs make the unglamorous parts visible: what data you are about to send, how much context it takes, how a document gets divided before retrieval, and what the manual hours cost. The browser-based tools do not send pasted text to a Rametric server.

Data handling

Scrub

Find obvious emails, phones, IBANs, AHV numbers, cards, JWTs and API-key patterns before text reaches a hosted model.

Runs in your browser

Token bench

Estimate the size of a prompt and compare rough context/cost implications before choosing a model.

Heuristic · not billing data

Split

Preview the windows a retrieval system might work with instead of treating a long document as one block.

Runs in your browser

Workflow

Automation savings

Hours per week, price per hour, the yearly cost of that work, and what a partial automation would take off the bill.

Runs in your browser

Applied examples

Prompt Forge

Tighten a prompt without changing what you meant.

Live

Second Mind

Answer, then critique, using your own OpenRouter key in the browser.

Live

CV Analyzer

Compare a CV with a job description and surface gaps.

Live

Local software

DocSumm

Local multi-format summarisation designed to sit beside a local model.

Code · GitHub

The Metric

A paid workflow product using structured career data and AI assistance.

Product · live

Want the private version?

The public labs demonstrate individual mechanics. A real internal system starts with your data boundary, workflow and approval points.

Talk through the workflow →