mlsanityData to 14 Oct 2026

From REST to MCP: An Empirical Study of API Wrapping and Automated Server Generation for LLM Agents

arXiv 2507.16044 ↗First posted 21 Jul 2025cs.SEcited by 14

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

The Model Context Protocol (MCP) is emerging as a standard interface through which LLM agents invoke external tools, and a growing ecosystem of MCP servers now mediates access to vendor services. Most of these servers target vendors that already expose REST APIs, yet the relationship between MCP tool interfaces and the underlying API surface has not been empirically characterised. This paper presents the first large-scale study of MCP server construction. We analyse 116 official servers to determine REST reliance and integration strategies (RQ1); examine servers paired with OpenAPI specifications to quantify operation exposure, omission, and mapping patterns (RQ2); evaluate automated generation from 80 real-world OpenAPI contracts (RQ3); and assess specification repair and tool-set transformations to improve correctness and reduce complexity (RQ4). We find that 88.6% of servers are fully or partially REST-backed, with 92% implementing tools as bare API wrappers. MCP servers expose a median of 19% of available operations, following systematic patterns predictable from the specification. Baseline generation succeeds for 76% of sampled tools; automated repair raises this to 94.2%, while filtering and regrouping reduce the median tool count per API by one-third. We release AutoMCP, an end-to-end pipeline integrating specification repair and empirically grounded tool-set transformations.

Key facts. Cited by 14 papers in the map (data through 14 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 19 of its 70 references that are in the map. First posted on arXiv 21 Jul 2025.

Builds on

19 of 70 references.

References · 19

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  3. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  4. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,947
  5. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
  6. 2022BERTopic: Neural topic modeling with a class-based TF-IDF procedurecited by 801
  7. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMscited by 591
  8. 2023RestGPT: Connecting Large Language Models with Real-World RESTful APIscited by 135
  9. 2024AnyTool: Self-Reflective, Hierarchical Agents for Large-Scale API Callscited by 124
  10. 2023T-Eval: Evaluating the Tool Utilization Capability of Large Language Models Step by Stepcited by 93
  11. 2025Model Context Protocol (MCP) at First Glance: Studying the Security and Maintainability of MCP Serverscited by 65
  12. 2025Retrieval Models Aren't Tool-Savvy: Benchmarking Tool Retrieval for Large Language Modelscited by 47
7 more
  1. 2016Understanding the Factors that Impact the Popularity of GitHub Repositoriescited by 27
  2. 2024An Empirical Study on Challenges for LLM Application Developerscited by 18
  3. 2025LongFuncEval: Measuring the effectiveness of long context models for function callingcited by 18
  4. 2024A Multi-Agent Approach for REST API Testing with Semantic Graphs and LLM-Driven Inputscited by 9
  5. 2025ToolFactory: Automating Tool Generation by Leveraging LLM to Understand REST API Documentationscited by 8
  6. 2021Restats: A Test Coverage Tool for RESTful APIscited by 4
  7. 2020Collecting Service-Based Maintainability Metrics from RESTful API Descriptions: Static Analysis and Threshold Derivationcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 9.

14 papers.

Sort:

By area:

  1. 2025Environment Scaling for Interactive Agentic Experience Collection: A Surveycited by 13
  2. 2025AutoTool: Dynamic Tool Selection and Integration for Agentic Reasoningcited by 11
  3. 2025Agentic Services Computingcited by 6
  4. 2026HarnessAPI: A Skill-First Framework for Unified Streaming APIs and MCP Toolscited by 1
  5. 2026SilentProbe: Measuring Silent Failure in Production APIs Used as Agent Toolscited by 1
  6. 2026Improving Proficiency and Efficiency of Android GUI Agents via Self-Generating Tool Actionscited by 1
  7. 2025HuggingR4: A Progressive Reasoning Framework for Discovering Optimal Model Companionscited by 0
  8. 2025Prompt-to-Parts: Generative AI for Physical Assembly and Scalable Instructionscited by 0
  9. 2026DADL: A Declarative Description Language for Enterprise Tool Libraries in LLM Agent Systemscited by 0
  10. 2026Octopus Protocol: One-Shot Hardware Discovery and Control for AI Agents via Infrastructure-as-Promptscited by 0
  11. 2026Delegating or Doing? Understanding User Behavior in Hybrid Human-Agent Interfacescited by 0
  12. 2026A Two-Dimensional Study of the Model Context Protocol: Publication and Adoptioncited by 0
2 more
  1. 2026MCP-GRANITE Benchmark: GRANularity Interface TEsting for MCP-Based LLM Agentscited by 0
  2. 2026MCP Error Messages Written for Developers Hurt the Most Capable Agents Mostcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Making OpenAPI Documentation Agent-Ready: Detecting Documentation and REST Smells with a Multi-Agent LLM System
  2. 2026DeltaMCP: Incremental Regeneration via Spec-Aware Transformation for MCP servers
  3. 2025A Measurement Study of Model Context Protocol Ecosystem
  4. 2026OOPS: Automated generation of REST API specification via LLMs ↗
  5. 2026An Empirical Study of Model Context Protocol Applications
  6. 2025A Framework for Testing and Adapting REST APIs as LLM Tools
  7. 2026Scalable LLM Agent Tool Access in the Cloud
  8. 2025From Specification to Service: Accelerating API-First Development Using Multi-Agent Systems