---
title: Reality Divergence
emoji: 😻
colorFrom: purple
colorTo: purple
sdk: gradio
sdk_version: 5.50.0
python_version: '3.11'
app_file: app.py
pinned: false
license: mit
tags:
  - build-small
  - backyard-ai
  - thousand-token-wood
  - track:backyard
  - track:wood
---

# Reality Divergence + Elsewhere

One shared reasoning engine, two lenses on the lives and worlds we didn't live.

- **🪞 Elsewhere (Your Timeline)** — a compassionate, safety-aware counterfactual
  reflection on a single personal choice. *(Track: Backyard AI)*
- **🌍 Reality Divergence (The World)** — speculative alternate history: pin a
  divergence point, reason forward through plausible ripples. *(Track: Thousand Token Wood)*

## Links

- 🎬 **Demo video:** https://youtu.be/cBzv7mwl_5E
- 📣 **Social post:** https://www.linkedin.com/posts/prashant-kumar-96460963_reality-divergence-a-hugging-face-space-activity-7472308555153391616-0Ub1

## Model rationale (Build Small)

- **Reasoning:** [`Qwen/Qwen2.5-7B-Instruct`](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
  — open weights, **well under the 32B cap**, served via the HF Inference API
  (`chat_completion`). Configurable through the `MODEL_NAME_TEXT` env var.
- **Rendering:** the model returns structured JSON, rendered as an animated HTML
  timeline (glowing year-node spine, branch cards, reflection prompts).
- **Atmosphere:** each result is topped with a **text-free** cinematic image
  ([`black-forest-labs/FLUX.1-schnell`](https://huggingface.co/black-forest-labs/FLUX.1-schnell)
  via `text_to_image`). Prompts explicitly forbid lettering, so no garbled "text"
  artifacts — just mood. Degrades gracefully if image generation is unavailable.
- **No GPU required to run** — all inference goes through the Inference API, so the
  Space stays light and within Zero GPU limits.

## Responsible UX

- **Elsewhere** runs input validation + risk detection and, on high-risk input,
  shows a safety notice with crisis resources instead of generating a timeline.
  Every timeline ends with an explicit **return-to-actual-life** strengths section.
- **Reality Divergence** is speculative and constrained to plausible, non-fantasy outcomes.
- The app **degrades gracefully**: model/timeout/JSON failures fall back to curated
  content so it never crashes.

## Run locally

```bash
pip install -r requirements.txt
cp .env.example .env   # then paste your HF token into .env
python app.py
```

Get a token at https://huggingface.co/settings/tokens (a *Read* token is enough).
Without a valid token the app still runs and shows fallback content.

## Submission checklist

- [x] Gradio Space, model under 32B, open weights
- [x] Dual-track UX (Backyard AI + Thousand Token Wood)
- [x] Safety gate + graceful fallback
- [x] Demo video link (see Links above)
- [x] Social media post link (see Links above)

---

Configuration reference: https://huggingface.co/docs/hub/spaces-config-reference