ScrapeGraphAI-100k: Dataset for Schema-Constrained LLM Generation

TL;DR

This paper introduces ScrapeGraphAI-100k, a large, real-world dataset of schema-constrained extraction events for training and benchmarking language models in structured output tasks.

Contribution

It provides a substantial, diverse dataset from real practitioner workloads, addressing limitations of synthetic and text-only datasets for schema-constrained generation.

Findings

A 1.7B fine-tuned model closely matches GPT-5-nano's output distribution.

The dataset covers 18,000+ schemas across 15 languages, with high coverage of English and Chinese.

Schema complexity impacts model performance, revealing sharp failure thresholds.

Abstract

Producing output that conforms to a specified JSON schema underlies tool use, structured extraction, and knowledge base construction in modern large language models. Despite this centrality, public datasets for the task remain small, synthetic, or text-only, and rarely pair real page content with the prompts and schemas used in practice. We introduce ScrapeGraphAI-100k, 93,695 schema-constrained extraction events collected via opt-in ScrapeGraphAI telemetry in Q2--Q3 2025, deduplicated and balanced by schema from 9M raw events. The corpus spans 18 000+ unique schemas across 15 named languages plus a long-tail Other category, with English and Traditional Chinese covering 88% of detected content, each instance pairs Markdown-converted page content with a prompt, schema, LLM response, and per-example jsonschema-rs structural conformance labels (semantic correctness is out of scope, and raw HTML is deferred beyond v1.0). We characterize structural diversity across the corpus and identify sharp failure thresholds as schema complexity grows. As a case study, a 1.7B student fine-tuned on this data closely tracks the output distribution of its GPT-5-nano teacher, though it still trails a 30B-A3B reference (3.3B active parameters) on schema compliance. We offer this distillation result as preliminary evidence that grounding schema-constrained generation in real practitioner workloads at scale enables training and benchmarking that prior synthetic or text-only corpora could not support.