The dataset covers all four released chapters of Deltarune. Each chapter is distributed as a separate pair of files (.txt and .jsonl) to allow selective loading.
The dataset spans 10K–100K records total across all four chapters. Each chapter is a self-contained file — you can load individual chapters without pulling in the entire dataset.
Summary table
| File | Chapter | Route coverage | Status |
|---|
chap1_cleaned.txt / .jsonl | Chapter 1 | Full (no alternate routes) | Stable |
chap2_cleaned.txt / .jsonl | Chapter 2 | Normal Route only | Stable |
chap3_cleaned.txt / .jsonl | Chapter 3 | Normal Route + Sword Route | Stable |
chap4_cleaned.txt / .jsonl | Chapter 4 | Normal Route only | Beta |
Chapter 1
Chapter 1 is fully transcribed and stable. This chapter has no alternate routes in the game, so there are no branching paths to account for.
Limitation: Chapter 1 was processed before the vid2text pipeline was in place. This means it lacks visual/stage direction descriptions that later chapters include for key scenes. No updates to address this are planned — use it as-is or exclude it if visual description consistency matters for your use case.
Chapter 2
Chapter 2 is fully transcribed for the Normal Route and is stable.
Limitation: Approximately 15 key scenes are missing visual/stage direction descriptions, which will be added in a future update. The Snowgrave/Weird Route — an alternate path with significant content differences — is not included. See Known gaps for details.
Chapter 3
Chapter 3 is fully transcribed and stable. This is the only chapter with alternate route content included: both the Normal Route and the Sword Route are transcribed.
Limitation: Approximately 15 key scenes are missing visual/stage direction descriptions, to be addressed in a future update.
Chapter 4
Chapter 4 is fully transcribed for the Normal Route and is currently in Beta status.
Chapter 4 may have quality issues compared to the stable chapters. The Snowgrave Route differences for Chapter 4 are not transcribed. Beta status means the transcription has not undergone the same level of quality control as Chapters 1–3.