story-iter and 1Prompt1Story

These are competitors—both tackle multi-image narrative generation from text, with story-iter using iterative refinement across frames while 1Prompt1Story achieves consistency through a single-prompt mechanism, requiring users to choose between the training-free iterative approach versus the prompt-engineering approach.

story-iter
65
Established
1Prompt1Story
49
Emerging
Maintenance 17/25
Adoption 10/25
Maturity 16/25
Community 22/25
Maintenance 6/25
Adoption 10/25
Maturity 16/25
Community 17/25
Stars: 949
Forks: 129
Downloads:
Commits (30d): 6
Language: Python
License: MIT
Stars: 313
Forks: 39
Downloads:
Commits (30d): 0
Language: Python
License: MIT
No Package No Dependents
No Package No Dependents

About story-iter

UCSC-VLAA/story-iter

[ICLR 2026] A Training-free Iterative Framework for Long Story Visualization

Implements a plug-and-play Global Reference Cross-Attention (GRCA) module that iteratively refines generated frames by incorporating all previous reference images during diffusion denoising, enabling semantic consistency across long sequences (up to 100 frames). Built on SDXL with IP-Adapter integration, the framework operates training-free and supports style control (comic, film, realistic) and ControlNet skeleton guidance for precise character pose management.

About 1Prompt1Story

byliutao/1Prompt1Story

🔥ICLR 2025 (Spotlight) One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation Using a Single Prompt

Leverages diffusion models with a novel consistency mechanism to generate multi-image sequences that maintain visual coherence across frames without requiring per-image prompts or fine-tuning. Built on PyTorch/Diffusers with Gradio interface support, it includes the Consistory+ benchmark for evaluating cross-image consistency and supports extended narrative generation from single text descriptions.

Scores updated daily from GitHub, PyPI, and npm data. How scores work