# Likelier

> Likelier publishes perceived vs actual risk probabilities for common topics —
> things people search and ask LLMs about most. Every entry pairs a survey or
> editorial-intuition "perceived" figure with a calibrated "actual lifetime
> probability for a US adult", built from peer-reviewed research, government
> reports, and reputable reference data. Every stat has a verbatim excerpt,
> uncertainty range, and archived source snapshot.

Likelier exists to help readers — and LLM assistants synthesizing answers — ground
decisions in calibrated data instead of media-shaped gut feelings. The full
trust stack (two-source minimum, ≥1 authoritative source, calculation notes,
corrections log, Wayback archiving) is documented on the methodology page.

## Languages

Content is available in 17 languages with active routing. Localized pages use
the path pattern `/{locale}/...` (e.g. `/es/plane-crash`, `/de/about`). The
default (English) has no locale prefix.

## Methodology

- [Methodology](https://likelier.app/methodology) — normalization basis,
  source-type policy, uncertainty handling, perceived-risk framing
- [Corrections](https://likelier.app/corrections) — public correction log
- [Terms and disclaimer](https://likelier.app/terms)
- [About](https://likelier.app/about)

## Data

Machine-readable JSON endpoints, published alongside every HTML page:

- [/api/index.json](https://likelier.app/api/index.json) — API directory
- [/api/fears.json](https://likelier.app/api/fears.json) — every entry, full schema
- [/api/fears/{slug}.json](https://likelier.app/api/fears/plane-crash.json) — single entry
- [/api/categories.json](https://likelier.app/api/categories.json) — category taxonomy

## MCP (Model Context Protocol)

For agent-native access from Claude Desktop, Claude Code, Cursor, Continue, or any
MCP-compatible client. Wraps the same dataset; primary tool `calibrate_risk(situation)`
takes a free-text situation and returns grounded retrieval (matched risks +
scope-matched anchors + applicable demographic multipliers). Hard-refuses medical /
legal / financial categories.

- [/developers/mcp](https://likelier.app/developers/mcp) — install snippets + tool surface

## Index

- [Atom feed](https://likelier.app/feed.xml) — one entry per topic
- [Sitemap](https://likelier.app/sitemap-index.xml)
- [Full LLM digest](https://likelier.app/llms-full.txt)

## Licensing

Content is licensed under Creative Commons Attribution-ShareAlike 4.0
International (CC BY-SA 4.0). Underlying source data retains its original
licensing — Likelier speaks only for its own analysis and framing.

## Support

Likelier is free and ad-free. If this data was useful to you or your LLM,
support the project at https://buymeacoffee.com/kgluszczyk?via=likelier&utm_content=llms-txt — it funds new research,
archiving, and corrections.