canonical: https://jentic.com/apis/amadeus.com/amadeus-airport-on-time-performance

# Amadeus Airport On-Time Performance

The Amadeus Airport On-Time Performance API predicts the probability that flights from a given airport will depart on time on a specified date. The single GET endpoint at /airport/predictions/on-time accepts an airport IATA code and a date and returns a probability score derived from historical performance and forecast conditions. It is used to flag risky connections and to decide whether to recommend buffer time on multi-leg itineraries.

## For AI agents

Predict on-time departure probability for a given airport and date. Agents call /airport/predictions/on-time to flag risky itineraries before booking.

## Scope

Does not return per-flight delays, live flight status, or arrival predictions - use for predicted on-time departures from a single airport on a single date only.

## Capabilities

- Predict the on-time departure probability for an airport on a future date
- Compare predicted reliability across alternative connecting airports
- Flag itineraries with high disruption risk before booking confirmation
- Recommend longer connection buffers when predicted on-time score is low
- Pull predictions for a planned travel date to inform itinerary advice

## Use cases

### Connection Risk Assessment

When a traveller is choosing between two itineraries, surface the predicted on-time performance of the connecting airport on the travel date. The /airport/predictions/on-time endpoint returns a probability score that can be displayed alongside layover duration, so the user knows whether a 60-minute connection at a low-reliability airport is sensible. This is more accurate than static airport rankings because the prediction reflects the specific date.

Example prompt: Call /airport/predictions/on-time for ORD on 2026-09-15 and report the on-time probability score.

### Travel Insurance Pricing

Travel insurers and trip-protection products can price disruption coverage using the predicted on-time score for the departure and connecting airports. A higher predicted reliability lowers the expected payout for missed-connection cover, which feeds directly into a dynamic premium. The single-endpoint design keeps the call cheap enough to run inline at quote time.

Example prompt: Fetch on-time predictions for FRA and IAD on 2026-10-04 and feed both scores into the premium calculation.

### Operations Buffer Planning

Corporate travel managers planning multi-leg itineraries can use predicted on-time scores to decide whether to recommend a longer ground-transfer buffer. A predicted reliability below a threshold triggers the booking tool to require at least 2 hours between flights, reducing disruption tickets opened with the support team.

Example prompt: Look up the on-time score for the user's connecting airport and require a 2-hour buffer if the score falls below 0.7.

### Agent Itinerary Advisor

An AI travel agent comparing two itineraries can call the prediction endpoint for each connecting airport to recommend the more reliable route. Through Jentic the agent searches for 'predict airport on-time performance', loads the schema, and runs the lookup against each candidate hub. The agent then summarises the trade-off between price and reliability for the user.

Example prompt: Use Jentic to call the on-time prediction endpoint for each of two candidate connecting airports and recommend the higher-scoring option.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | /airport/predictions/on-time | Predict on-time departure probability for an airport |

## Key resources

- **On-Time Prediction** — GET predicted on-time departure probability for an airport on a date.

## Why Jentic

- **Setup:** Wiring the Airport On-Time Performance API by hand means running the Amadeus OAuth 2.0 client_credentials exchange, refreshing the short-lived bearer token yourself, and pointing at the correct Amadeus host and version before the single GET /airport/predictions/on-time call will run. Through Jentic you install once, import Airport On-Time Performance from the API Directory, store the Amadeus client credentials once, and your agent calls it.
- **Permission scoping:** This API carries the airport IATA code and date as query parameters, not as a resource id in the URL path, so scoping is by operation rather than by resource. You limit the agent to the operations it needs, such as reading the on-time departure prediction, and it cannot reach other Amadeus operations unless you add them.
- **Credential handling:** Your Amadeus client_id and client_secret are stored once, encrypted, by your own Jentic One instance, which runs the client_credentials exchange and injects a fresh bearer token at execution time. The raw secret never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'check whether flights from JFK are likely to depart on time tomorrow', and Jentic returns the /airport/predictions/on-time operation with its input schema so the agent supplies the airport code and date without reading the Amadeus docs first.

## Related APIs

- **Amadeus Flight Delay Prediction** — Predicts delay risk for a specific flight rather than a whole airport.
- **Amadeus On-Demand Flight Status** — Returns the live status of a specific flight after departure date is near.
- **Amadeus Flight Availabilities Search** — Surfaces flight options whose connecting airports can be scored for reliability.

## FAQ

### What authentication does the Airport On-Time Performance API use?

The API uses the Amadeus OAuth 2.0 client-credentials flow. Exchange API key and secret at the Amadeus token endpoint and send the bearer token on the /airport/predictions/on-time request. Jentic keeps the credentials in the encrypted vault and refreshes the token automatically.

### What does the on-time probability score represent?

The score is the predicted likelihood that flights departing the supplied airport on the supplied date will leave on time. It is a forward-looking estimate derived from historical performance, not a record of past flights, so it is most useful for itineraries that have not yet been flown.

### What are the rate limits for the Airport On-Time Performance API?

Rate limits are not declared in the OpenAPI spec and depend on the Amadeus contract tier. The endpoint is read-only and cacheable for the same airport and date, so cache aggressively to stay under the per-second cap for your account.

### How do I check on-time risk for a flight through Jentic?

Run pip install jentic, then use the search query 'predict airport on-time performance'. Jentic returns the /airport/predictions/on-time operation, you load the schema, supply airportCode and date, and execute. The score can then drive a recommendation in your itinerary UI.

### Can I get a per-flight prediction or only per-airport?

This endpoint is per-airport per-date - it predicts overall on-time performance for departures from the supplied airport. For per-flight delay predictions Amadeus offers a separate Flight Delay Prediction API that takes a specific carrier, route, and departure time.

### Is the Airport On-Time Performance API free?

Amadeus self-service APIs include a free test environment with quota limits and a paid production tier. Production pricing depends on transaction volume and is set by contract - confirm current rates in the Amadeus self-service developer portal.

### Can I limit what my agent is allowed to do with the Airport On-Time Performance API?

Yes. Because you run Jentic One yourself, your own rules decide which operations the agent may call and which credentials it may use. This API takes the airport IATA code and date as query parameters rather than in the URL path, so scoping is by operation: you can allow the agent only the GET /airport/predictions/on-time read and nothing else. The agent cannot reach any other Amadeus operation unless you explicitly add it.