Model card: add proper LTX-2 attribution, license section, commercial-use notice

README.md

@@ -2,6 +2,8 @@
 language:
   - en
 license: other
+license_name: ltx-2-community
+license_link: https://huggingface.co/ResembleAI/Dramabox/blob/main/LICENSE
 pipeline_tag: text-to-speech
 tags:
   - tts
@@ -11,6 +13,8 @@ tags:
   - flow-matching
   - ltx-2
 library_name: ltx-audio-tts
+base_model: Lightricks/LTX-2
+base_model_relation: finetune
 ---
 
 # Dramabox - Expressive TTS with Voice Cloning
@@ -141,6 +145,18 @@ The prompt is a scene description that controls how the model speaks:
 | Warm server (recommended) | **~24 GB** | **~2.5s** |
 | Cold inference (per-call loading) | ~8 GB peak | ~30s |
 
-## License
+## License & Attribution
 
-Built on [LTX-2.3](https://github.com/Lightricks/LTX-2) by Lightricks. Please refer to the LTX-2 license for usage terms.
+Dramabox is a fine-tuned derivative of **LTX-2** by Lightricks Ltd. (https://github.com/Lightricks/LTX-2) released on January 5, 2026. Under the LTX-2 Community License Agreement, the weights and code in this repository are **Derivatives of LTX-2** and are distributed under the **same license** as the base model.
+
+Full license text: [`LICENSE`](https://huggingface.co/ResembleAI/Dramabox/blob/main/LICENSE) (LTX-2 Community License Agreement, unchanged from the upstream release).
+
+### Key points for users
+
+- **Free use** is permitted for individuals and entities with **annual revenue below US $10,000,000**, subject to the [use restrictions in Attachment A](https://github.com/Lightricks/LTX-2/blob/main/LICENSE) of the license (including prohibitions on harmful content, deepfakes without consent, military use, etc.).
+- **Commercial Entities** with **annual revenue ≥ US $10,000,000** must obtain a paid commercial license directly from Lightricks (https://ltx.io/model/licensing) before using LTX-2 or any Derivative, including this model.
+- **Competing products:** per Attachment A §20, LTX-2 and its Derivatives may not be used in products that directly compete with Lightricks' commercial offerings without a separate commercial license.
+- **Redistribution:** any redistribution of this model — modified or unmodified — must include a copy of the LTX-2 Community License and be distributed exclusively under its terms. You may not sub-license this model under more permissive terms (e.g., MIT, Apache).
+- **Modifications to the upstream model:** voice-cloning adapter weights were trained by Resemble AI and merged into the DiT checkpoint published here. Inference code in this repository is adapted from the upstream LTX-2 release with additions for the voice-cloning pathway.
+
+Not legal advice. Users are responsible for reviewing the full license and confirming their own compliance before use.