About OpenAirDocs
OpenAirDocs is an independent project to aggregate and organize public aviation data. This data is collected from various APIs and websites of U.S. federal sources.
Our mission is to support a safe and thriving aviation ecosystem.
Our objective with OpenAirDocs is to make aviation data more accessible to human readers, as well as AI models. We publish LLM-friendly data definitions, and our Parquet databases can be queried without download.
This site is a project of Contrail Labs, and all the data is released into the public domain. We’d consider this the CC0 1.0 Universal Deed. Future data sources may require attribution, due to upstream.
We intend to expand to other international data.
We intend to open source the code shortly.
If you have feedback or questions about this data, please email andy@contraillabs.com.
Where the data comes from
Every record starts as a public file from a U.S. federal agency. OpenAirDocs downloads it, keeps a copy and republishes it with the agency’s own fields and wording.
- FAA Dynamic Regulatory System: the AD index and statuses, the products each AD names, older AD text, proposed ADs, SAIBs, concern sheets, advisory circulars, type certificate data sheets and STCs.
- Federal Register: the published text of ADs and proposed ADs from 1994 on.
- FAA aircraft registry: registered aircraft, their owners and cancelled registrations.
- FAA service difficulty reports: what operators and mechanics reported going wrong, from 1995 on.
- FAA Accident/Incident Data System: accident and incident reports from 1973 on.
- NTSB: investigations from 1962 on, from the NTSB’s public API and its Access database files, with a list of each case’s docket files.
- Public-domain place lists (FAA airports, Census towns and ZIP areas, USGS place names, OurAirports), used only to put records on a map.
Each AD page lists the files it was built from, with their SHA-256 checksums. The status page shows when each source was last read.
What OpenAirDocs adds
Most of what you see is the agency’s record as published. Every column in the data has one of four labels, so a match we made is never mistaken for an FAA or NTSB finding:
- source
- The agency’s own words and numbers. At most, a date, number or yes/no is converted.
- normalized
- The agency’s value under a fixed rule, such as a code spelled out or the N added to a registration number.
- derived
- Inferred by OpenAirDocs: an AD matched to aircraft models, an applicability paragraph pulled from an AD’s text, an accident report matched to a registered aircraft, a point placed on a map. These can be wrong.
- openair
- Identifiers, links and download details we add. They say nothing about an aircraft or an event.
Derived values sit beside the agency’s record and never replace it. On this site they carry a dashed ring, and in the Markdown copies “(OpenAirDocs)”.
For technical readers: every column’s label and description are in the data dictionary, in manifest.json, in the openair_column table of each database and in each Parquet field’s metadata. The release README describes each source, source_file lists every file we copied with its original address and SHA-256, and the downloads list every file in the release.
How records are matched
Each registered aircraft has an airframe model code and an engine model code. OpenAirDocs matches those codes to the makes and models each AD names, and to the models on each type certificate data sheet. A match means the AD may apply. An AD’s applicability can narrow it further by serial number, part number or installed equipment, and ADs on appliances and propellers aren’t matched by model. The lists are a starting point for research, not a compliance determination.
NTSB and FAA accident reports are matched to registered aircraft by N-number and date, which can go wrong when a number has been reused. Map points use the NTSB’s own coordinates when they fall in the state it names, or else the named airport or town. An owner’s point is their mailing ZIP area, not where the aircraft is kept.
What is official
Only the Federal Register edition of an AD, published on govinfo.gov, is the official legal text, and the NTSB’s own report is the official record of an investigation. Most record pages link to the agency’s own page or file; check anything important there. OpenAirDocs is not affiliated with the FAA or the NTSB. Source documents are U.S. government works in the public domain.
Using the data
The release is published as Parquet, DuckDB and SQLite under CC0 1.0; see Use the data and Collections. Add .md to any page’s address for a Markdown copy.
Privacy
Aircraft pages show the registered owner’s name, city and state, as the FAA registry publishes them. Street addresses are left out. Owners who asked the FAA to withhold their information are not in the FAA’s public file, so they are not here either.