Strata
Strata is a living meta-analysis like tool for the teams who depend on evidence.
Core trick
Claude extracts trial effect data and citations, a statistics library pools them, and rule-based checks flag implausible extractions before pooling.
Description
Strata is a living meta-analysis like tool for the teams who depend on evidence. A meta-analysis can pull scattered trial data into a single defensible answer, but it goes stale the day it publishes, and refreshing one by hand is slow and expensive. Strata keeps that answer current and lets you trace every number back to its source.
You ask a research question, and Strata pulls the matching trials, reads the results out of each one, and gives a first read on how much to trust each. A person stays in the loop: Strata surfaces low-confidence extractions and the trust judgments that carry weight for a human to confirm before anything is pooled, rather than deciding those alone. It then combines what passes into a single answer: the size of the effect, how confident you can be, a forest plot, and a plain-English summary, with every figure clickable back to the sentence it came from. When a new trial reports, it re-runs and tells you whether the answer changed. This works because of one rule: Claude reads, code does the maths. Claude finds and structures the numbers; a statistics library does all the pooling; and plain sanity checks flag anything that doesn't add up rather than folding it in. When the data is too thin to answer, Strata says so instead of inventing a confident number. Claude drives the whole workflow through an MCP server, with a full web platform and CLI over the same core.
README excerpt
# Strata Strata is a living meta-analysis tool. You ask a clinical question in PICO form, and it finds the trials, extracts the effect data, appraises it, and pools it into one answer: an effect estimate with a confidence interval, a forest plot, heterogeneity measures, and a plain-language summary. Every number links back to the trial and the snippet it came from. When a new trial reads out, Strata re-runs and tells you whether the estimate or the conclusion changed. The problem it addresses is that a meta-analysis can go stale soon after the day it is completed, and updating one by hand is slow and expensive. https://livemeta-backend-production.up.railway.app/ ## How it divides the work The design splits the pipeline by what each part is good at. The LLM does the reading. It parses the question into PICO, pulls arm-level results out of trial records, and does a first-pass RoB 2 and GRADE read. It never does arithmetic on the results. Deterministic code does the numbers. All pooling runs through a validated statistics library (random-effects, REML, HKSJ intervals). Before anything is pooled, a validation gate checks it: events can't exceed arm totals, arm sizes have to sum, percentages have to match counts. Whatever fails that check is flagged for review instead of pooled. Two other rules fall out of this. Every extracted value and every risk-of-bias judgment carries its source trial ID and the exact sentence or table cell behind it. And when the data is thin or het