# vaers.globalpatientsafety.com

> VAERS vaccine × event disproportionality. Hypothesis, not causation.

Public aggregates from CDC VAERS. Read `n` and `eb05`. No API key.
Live extract through-quarter: `X-VAERS-Through` on GET /signals (also
`vaers_through` on `format=profile`, `format=series`, and `format=class`,
and `Data through` on `format=brief`).

VAERS reports are voluntary. Counts are reporting, not incidence.
Disproportionate reporting is a hypothesis, not evidence of causation.

## Docs

- [API.md](https://vaers.globalpatientsafety.com/API.md): human REST docs for GET /signals
- [OpenAPI](https://vaers.globalpatientsafety.com/openapi.json): OpenAPI 3 schema for GET /signals (json / csv / series / profile / class / brief)
- [ChatGPT Actions](https://vaers.globalpatientsafety.com/chatgpt.md): paste-ready Custom GPT import of that schema
- [Signals API](https://vaers.globalpatientsafety.com/signals?vaccine=COVID19&event=Myocarditis&format=brief): example pair brief
- [Event profile](https://vaers.globalpatientsafety.com/signals?event=Myocarditis&format=profile): compact JSON for one event
- [Human UI](https://vaers.globalpatientsafety.com/): search and selected-pair series table
- [MCP](https://faers.mobi/mcp): tools `get_vaers`, `vaers_search_vaccine`, `vaers_get_signal`, and `vaers_get_profile` on the existing faers.mobi MCP server. This host does not serve /mcp.

`n` is the sum of quarterly observed counts. Those windows overlap, so `n` is not a count of unique reports. `n_unique` is not in this extract. `eb05` is the peak quarterly EB05 (the app's Peak EB05). `eb05_peak_pre2021` is the highest quarterly EB05 before 2021Q1 on the series. `peak_in_covid_era` is true when the vaccine is not COVID19 and that series peak falls in 2021Q1–2022Q4. `background=no_covid` is not served. `ewma_eb05` is separate. There is no FDA-label, substance, or PDF route on this host. `format=label` and `format=pdf` return HTTP 400.