Overview
Discovery API
Developers
Shopping Intelligence
Case studies
Available now

Reduce travel API calls without brute-force shopping.

MyEscapePlan is a pre-shopping decision layer designed to reduce unnecessary downstream travel API calls. It ranks destinations, date windows and trip lengths before live flight, hotel or package inventory is requested.
MyEscapePlan does not replace inventory providers. It provides the travel-specific planning and pruning layer in front of them.
Discovery itself is planner-only and makes no live supplier calls. Provider-backed Verified Search is a separate optional workflow for pilots that need it.
Built for AI travel agents, conversational travel products and search platforms where one traveller request can branch into many downstream shopping decisions and provider API calls.
From London, 4 days next month, under £300pp, somewhere warm with direct flights.
Example constrained shopping plan
1
Málaga
3–5 days · flexible window
Strong fit
2
Alicante
3–4 days · flexible window
Strong fit
3
Palma
3–5 days · verify fare first
Possible fit
Independent industry evidence

Travel shopping already operates at volumes where fewer API calls matter.

IATA's 2019 scalability study reported OTA and metasearch look-to-book ratios of roughly 1,000–10,000+:1 shopping requests per booking. Agentic travel adds another source of iterative, automated search as AI travel agents explore and refine options on a traveller's behalf.

1,000–10,000+:1

OTA & metasearch look-to-book
IATA's 2019 scalability study reported roughly 1,000:1 in best cases and commonly above 10,000:1 for OTA and metasearch shopping.
IATA NDC Scalability Study · 2019 ↗

100–300:1

Airline-direct look-to-book
IATA's 2019 scalability study contrasted OTA and metasearch volumes with roughly 100–300 shopping requests per booking on airlines' own websites.
IATA NDC Scalability Study · 2019 ↗

150+

Tool calls in long-horizon agentic travel
TRIP-Bench reports realistic long-horizon travel dialogues that can involve more than 150 tool calls across 18 travel tools.
TRIP-Bench · 2026 ↗
Look-to-book is an industry shopping-to-booking ratio, not a claim that one traveller request creates 1,000–10,000 HTTP calls. TRIP-Bench separately reports 150+ tool calls in realistic long-horizon travel dialogues; those calls span travel tools and are not flight-supplier calls alone.
IATA 2025 Look-to-Book ↗TRIP-Bench ↗TREK travel-agent benchmark ↗
Why reducing travel API calls matters

At 1,000–10,000+:1 look-to-book, small reductions compound.

The IATA range above describes shopping requests per booking, not API calls from one traveller request. As an illustrative example, reducing shopping volume by 10% at that industry baseline would mean roughly 100–1,000+ fewer shopping requests per booking.
Where each shopping request triggers one or more downstream provider API calls, reducing the shopping space can reduce downstream API traffic further. Actual savings depend on the customer's architecture, provider contracts and achieved pruning rate.
Potential net saving = avoided downstream provider calls × marginal provider cost − MyEscapePlan cost.
This is an illustration of the economics, not a measured MyEscapePlan performance claim. Published MyEscapePlan reduction figures will appear only after the audited benchmark is available.
What the API does

Reduce search fan-out before live inventory calls.

Destination candidate generation

Resolve open-ended intent against a managed destination universe while preserving hard traveller constraints.

Intelligent date and duration selection

Reason across flexible windows and trip lengths to prioritise useful combinations instead of forcing one brittle interpretation or shopping every possibility.

Ranked shopping candidates

Prune, rank and explain the combinations worth live verification, with controlled relaxation when exact options fail.
Search-space intelligence

Not every plausible date deserves a supplier call.

A flexible request can imply hundreds of destination, date and duration combinations. MyEscapePlan narrows that space first, prioritising combinations using traveller constraints, flexibility, affordability signals and candidate fit.
Why now · Agentic travel

We expect agentic travel to multiply shopping volume.

As agentic travel matures, AI travel agents can branch across destinations, dates, durations, airports, budgets and follow-up refinements from a single user request — then re-check the most promising options. Every branch can create downstream flight, hotel or package shopping work.
IATA's 2025 Look-to-Book work explicitly highlights Generative and Agentic AI when discussing rising search volumes and offer creation. MyEscapePlan is designed to sit before expensive live inventory APIs and decide which branches are worth shopping.
01

One traveller request

A human can ask broadly: somewhere warm, four days next month, under budget, preferably direct.
02

An agent explores many branches

Agentic travel planning can autonomously test destinations, dates, durations, airports and revised constraints instead of waiting for manual clicks.
03

Discovery narrows live shopping

MyEscapePlan ranks and prunes candidate combinations first, so downstream flight-shopping APIs receive the small set most worth verifying.
Benchmarking principle
Recent travel-agent benchmarks such as TREK treat tool efficiency as a first-class evaluation dimension alongside task correctness. MyEscapePlan applies the same principle to shopping: preserve the good options while reducing unnecessary downstream work.
Read TREK ↗
Why this layer exists

Supplier APIs answer “what inventory exists?” They do not decide the entire travel search strategy.

Discovery requires intelligent selection across destinations, dates and durations, plus affordability reasoning, candidate pruning and fallback rules before high-cost live shopping begins.
Natural-language intent parsing
Managed destination universe and customer-provided destination allowlists
Intelligent date-window and duration selection
Affordability signals
Candidate pruning and ranking
Fallback and relaxation logic
Planner-only discovery with no live supplier calls
Optional provider-backed Verified Search as a separate workflow
Reasons and qualitative fit for downstream UX
Best-fit customers

For teams building travel products, not another inventory source.

AI-native travel

Trip-planning startups, conversational travel assistants, itinerary builders and travel-agent copilots.

Existing travel platforms

OTAs, metasearch, package-holiday companies, flight-search products and publishers adding discovery.

Agent and LLM platforms

General AI products that need travel discovery capability without rebuilding travel-specific planning infrastructure.
Discovery API Pilot

4–6 week integration pilot

API access, a generous fixed usage allowance, integration support, basic usage analytics and enough real traffic to evaluate recommendation quality and shopping-space reduction.

MyEscapePlan Travel Intelligence

Travel discovery infrastructure for reducing search fan-out, plus travel shopping data and market intelligence.
MYESCAPEPLAN LTD · Company no. 16395758 · Registered in United Kingdom
Discovery API is available now. Destination & Market Shopping Intelligence is coming soon.
Shopping Intelligence case studies will publish when the benchmark and methodology are ready for public release.
Legal
Terms & conditions
Privacy
Cookies