
Amadeus is phasing out its Self-Service portal in July 2026, forcing agencies on that tier to pick a new integration partner right now, not eventually. Here's how to actually make that decision.
Travel API decisions used to be relatively low stakes; most providers offered similar coverage, and switching later wasn't a major disruption. That's changed. With Amadeus's Self-Service shutdown forcing a real migration decision, NDC content splitting from traditional GDS fares, and provider financial stability becoming a genuine risk factor, the choice you make this year is more likely to shape your platform's reliability and cost structure for the next several years than it would have a decade ago.
The biggest mistake in this decision is treating "travel API provider" as one category. In practice, there are at least five distinct types, and each solves a different problem.
GDS platforms like Amadeus, Sabre, and Travelport give broad airline content and mature servicing workflows, ticketing, refunds, exchanges, built over decades. Our comparison of how the major GDS platforms differ covers regional strength and API quality across all three in more depth.
NDC-first providers, including newer entrants positioning themselves as alternatives to traditional GDS access, focus specifically on richer, airline-controlled offers, branded fares, dynamic pricing, ancillaries, that traditional GDS fare filing doesn't fully expose.
Marketplace partner APIs, common in the tours and activities space, give access to a specific marketplace's inventory, with a choice between owning the full booking flow or using a lighter referral integration that trades revenue per booking for simplicity.
Wholesalers and aggregators combine content from multiple suppliers into a single feed, aggregating GDS inventory across sources into one connection point, trading some margin for meaningfully faster integration.
Direct supplier and infrastructure platforms give the most control, and the most responsibility, letting a business manage its own scaling and infrastructure decisions rather than depending entirely on a third party's architecture.
This decision increasingly comes down to whether your business needs broad, standardized fare access or richer, airline-specific offers. Our full breakdown of the difference between GDS and NDC content covers this in detail, but the short version for a partner selection decision is this: if your business relies on broad multi-airline coverage and mature servicing tools, GDS access remains essential. If you sell heavily on airlines with strong NDC programs, a partner with solid NDC content access will likely save meaningfully on distribution surcharges and unlock content a GDS-only connection won't show you.
Most growing travel businesses need both, which means the real evaluation question isn't "GDS or NDC," it's whether a given partner handles both well, or whether you'll need two separate relationships to get full coverage.
Beyond the obvious criteria of price and content coverage, a few less obvious factors consistently separate a reliable partner from a risky one.
Documentation and sandbox quality. Some providers offer interactive API explorers, clear versioning, and working sample code. Others hand over a static PDF and little else. This directly affects your integration timeline and your ongoing maintenance cost, and it's something your development team can test firsthand before signing anything, not something you need to take on faith. Our own breakdown of common integration challenges agencies run into covers several of the documentation and certification gaps that show up most often in practice.
Sandbox behavior does not guarantee production behavior. This is a genuinely underdiscussed risk. One documented case involved a property management system partner whose sandbox environment updated availability every 60 seconds, but whose production environment only refreshed every 5 minutes during peak load, a gap that led directly to overbooked rooms and support chaos. Testing in a sandbox is necessary, but it isn't sufficient; ask specifically about production refresh rates and peak-load behavior, not just sandbox specs.
Financial stability and vendor lock-in risk. Research on vendor evaluation has found that 79 percent of businesses have lost measurable revenue due to supply chain disruptions tied to vendor instability over the past three years. A partner's financial health, ownership structure, and acquisition risk matter as much as their feature list, since a partner that gets acquired or restructured, as Amadeus's own Self-Service customers are now experiencing, can force an unplanned migration on your timeline, not theirs.
Support responsiveness, tested, not assumed. The pattern to watch for is a partner that's consultative and responsive before a contract is signed, then slow afterward. Ask for a named account contact, a contractually defined response SLA, and a willingness to share performance metrics rather than just sales promises, and test that responsiveness during evaluation, before you're locked in.
Pricing structure and total cost of ownership. Providers price differently, subscription fees, per-transaction charges, or revenue-sharing arrangements, and the headline rate rarely tells the whole story. Our breakdown of how GDS API pricing typically works is a useful reference point for understanding where hidden costs tend to show up.
A thorough evaluation, covering research, demos, reference calls, and real sandbox testing, typically takes four to eight weeks. Enterprise-scale integrations, involving full certification and contract negotiation, often run three to six months. Rushing this process is where most bad partner decisions come from, a structured, weighted comparison across cost, technical capability, and support quality consistently outperforms a decision made after one impressive demo.
To see what a well-executed integration looks like in practice, our Amadeus Quick Connect booking engine case study is a useful reference point. The project prioritized rapid implementation without sacrificing the reliability a growing booking platform depends on, exactly the balance most businesses are trying to strike when evaluating a new partner: how fast can this actually go live, and will it still hold up once real booking volume hits it.
Choose a primary GDS relationship if: your business depends on broad multi-airline coverage, complex servicing (refunds, reissues, exchanges), or established corporate travel workflows, where GDS platforms remain the most mature option available.
Prioritize NDC-first access if: a meaningful share of your bookings run through airlines with strong NDC programs, where the richer content and lower distribution surcharges deliver a measurable, direct cost advantage.
Consider an aggregator or wholesaler if: you need to launch quickly and can accept a modest margin tradeoff for faster time to market, particularly useful for B2B portal development where speed to a working platform often matters more than owning every supplier relationship directly.
Most growing businesses end up with more than one relationship rather than a single partner, and that's a normal, expected outcome, not a sign the first choice was wrong.
A travel API integration partner is a company or platform that provides programmatic access to travel inventory, flights, hotels, cars, or activities, letting a business search, book, and manage reservations without building direct connections to every individual supplier.
Start by identifying what you're building, an OTA, a B2B portal, a corporate tool, then evaluate providers on documentation quality, real production reliability (not just sandbox specs), financial stability, support responsiveness, and total cost of ownership, not just headline pricing or content volume.
Amadeus is phasing out its Self-Service portal in July 2026, requiring agencies on that tier to migrate to Amadeus Enterprise or choose an alternative provider, making partner selection an active, time-sensitive decision for affected businesses right now.
Neither is universally better, they serve different needs. GDS offers broad coverage and mature servicing tools. NDC offers richer, airline-specific content and typically lower distribution surcharges. Most growing travel businesses need both rather than choosing one exclusively.
A thorough vendor evaluation typically takes four to eight weeks, while enterprise-scale integrations, including certification and contract negotiation, commonly take three to six months.
Beyond price and coverage, look closely at documentation and sandbox quality, real production performance under peak load, financial stability and acquisition risk, support responsiveness before and after signing, and the true total cost of ownership.
Choosing a travel API integration partner in 2026 isn't a one-time decision you can make and forget, Amadeus's own Self-Service shutdown proves that even established, trusted providers can force an unplanned migration with little warning. The businesses that navigate this well are the ones that evaluate partners on production reliability and financial stability, not just feature lists and demo polish, and that build relationships flexible enough to add a second provider when one alone can't cover everything they need.
If you're evaluating what a migration or new integration would actually take for your platform, you can hire a travel technology developer to scope the specific work involved.
Sources: Amadeus, Self-Service API migration guidance, 2026; IATA, New Distribution Capability Program; vendor evaluation research on financial stability and supply chain disruption risk, 2026.