Agentic Coding Tool/Question/GraphQL vs REST API

Properties

Question: Are agentic coding tools better at using GraphQL vs REST APIs?

This is an open question about the comparative effectiveness of agentic coding tools when working with GraphQL APIs versus REST APIs.
Considerations
- GraphQL advantages for AI agents:
  - Strongly typed schema could help with tool/function calling
  - Single endpoint might simplify agent decision-making
  - Schema introspection provides self-documenting API structure
  - Queries allow requesting exactly needed fields
- REST API advantages for AI agents:
  - Multiple endpoints might provide clearer action boundaries
  - More familiar pattern - large language models trained on more REST examples
  - Standard HTTP verbs (GET, POST, PUT, DELETE) align with common agent actions
  - Typically more verbose documentation and examples available
- GraphQL challenges for AI agents:
  - Query construction requires understanding nested relationships
  - Mutations can be complex to construct correctly
  - May require understanding of GraphQL-specific concepts
- REST API challenges for AI agents:
  - May require multiple API calls to get related data
  - Less self-documenting - requires understanding full API structure
  - Over-fetching or under-fetching data

Title	Description	Year / Venue	Link
Sequential API Function Calling Using GraphQL Schema (GraphQL‑RestBench)	Dataset of natural‑language utterances paired with sequences of REST API calls using a GraphQL schema prompt to study LLM function‑calling.	2024 (EMNLP)	https://aclanthology.org/2024.emnlp-main.1083.pdf
PrediQL: Automated Testing of GraphQL APIs with LLMs	Retrieval‑augmented, LLM‑guided fuzzer for GraphQL APIs that uses adaptive bandit learning for query generation and vulnerability detection.	2025 (arXiv)	https://arxiv.org/html/2510.10407v1
Large Language Models – Based Fuzzing Techniques: A Survey	Survey of LLM‑based fuzzing tools and methods, including API/fuzzer contexts.	2024 (arXiv)	https://arxiv.org/html/2402.00350v1