War-craft improves web api recommendations by auditing requirements
WAR-CRAFT: Refining Web API Recommendations through Requirement Auditing
Artificial Intelligence
Summary
Web API recommenders suggest sets of APIs to match a user's needs but often include irrelevant or missing APIs because they don't fully check against official documentation. The authors created WAR-CRAFT, a method that reviews and improves those suggestions by removing unsupported or redundant APIs and adding missing ones based on documented provider capabilities. It uses a large language model and a fine-tuned component to audit and fix recommendations, resulting in more accurate and less redundant sets of APIs. This approach works well across different recommendation methods and keeps the set size balanced.
What this means in practice
- •For web service developers: Improve API mashup tool accuracy by refining recommended API sets to better match user requirements with verified provider documentation.
- •For software integrators: Reduce redundant or unsupported APIs in integration workflows by auditing API recommendations against official documentation automatically.
Authors
Zishuo Xu, Dezhong Yao, Zhiqiang Xu
Abstract
Web API recommenders return a set of APIs for a mashup request but do not check it against provider documentation, the main public evidence of an API's capabilities. The set can therefore contain unsupported or redundant APIs and miss requested capabilities. Filling every apparent gap does not fix this, because a gap may already be served, may not need its own API, or may have no documented provider. We study post-recommendation refinement, which revises a frozen recommender's output using only the request, the catalog, and provider documentation. We propose WAR-CRAFT, which treats refinement as three decisions. An LLM removes unsupported or subsumed APIs by judging candidates jointly, then audits the remaining set against a requirement checklist derived from the request. For each uncovered requirement, a fine-tuned catalog-constrained Resolver proposes a catalog API or abstains, and an independent documentation check admits only supported proposals. On ProgrammableWeb with 7 heterogeneous upstream recommenders, WAR-CRAFT raises mean SetF1 from 0.558 to 0.644, above the strongest of 5 refinement baselines (62.5%), and roughly halves functional redundancy (0.127 to 0.065) without changing mean functional coverage. Gains range from +2.2 points on generative recommenders to +18.2 on fixed-cutoff rankers. At matched set sizes, WAR-CRAFT still outperforms truncating the upstream output, so the gain is not explained by set-size calibration alone. Our code is available at https://github.com/ZsXu-enrico/WAR-CRAFT.