Spaces:
Running
Running
Align TensorScale organization card strictly with official website copy
Browse files
README.md
CHANGED
|
@@ -9,18 +9,18 @@ pinned: false
|
|
| 9 |
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
-
#
|
| 13 |
|
| 14 |
-
### **Powering the
|
| 15 |
|
| 16 |
*The AI stack was built for language models. We’re rebuilding it for world models.*
|
| 17 |
|
| 18 |
[](https://tensorscale.io)
|
| 19 |
[](https://console.tensorscale.io)
|
| 20 |
-
[](https://tensorscale.io/pricing.html)
|
| 22 |
[](https://tensorscale.io/blog.html)
|
| 23 |
-
[](https://x.com/TensorScaleAI)
|
| 25 |
[](https://www.linkedin.com/company/tensorscale-ai)
|
| 26 |
|
|
@@ -28,36 +28,43 @@ pinned: false
|
|
| 28 |
|
| 29 |
---
|
| 30 |
|
| 31 |
-
## About
|
| 32 |
|
| 33 |
-
AI is evolving from understanding and generating text to understanding and simulating the world itself—its geometry, physics, causality, dynamics, and interactions.
|
| 34 |
|
| 35 |
-
|
| 36 |
|
| 37 |
-
|
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
---
|
| 40 |
|
| 41 |
-
##
|
| 42 |
|
| 43 |
-
|
| 44 |
-
- **
|
| 45 |
-
- **
|
| 46 |
|
| 47 |
---
|
| 48 |
|
| 49 |
-
##
|
| 50 |
-
|
| 51 |
-
| Modality | Hosted Models | Capabilities | Delivery Modes |
|
| 52 |
-
| :--- | :--- | :--- | :--- |
|
| 53 |
-
| **Video + Audio & World** | **MiniMax H3**, **LTX-2.5 Fast**, **Wan 3.0**, **Cosmos3 Nano** | Text-to-Video+Audio (`T2VA`), First/Last-Frame (`FL2VA`), Multi-Reference (`Ref2VA`), 2K/4K Upscale & Video Editing | Synchronous, Fast-Start MP4 Streaming, Async Jobs |
|
| 54 |
-
| **Image Generation & Editing** | **MiniMax H3 Image**, **SenseNova U1.5**, **Qwen Image 2.1**, **MOSS Image** (Base & SFT), **FLUX3 Image Klein**, **Hunyuan Image 3 Fast** | High-density Text-to-Image, Multi-Reference Image Editing, Layout-Compiled Typography & Restoration | Synchronous, Async Jobs |
|
| 55 |
|
| 56 |
-
|
| 57 |
|
| 58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
-
|
| 61 |
|
| 62 |
```bash
|
| 63 |
curl "https://api.tensorscale.io/v2/ltx-2.5/fast" \
|
|
@@ -76,19 +83,48 @@ curl "https://api.tensorscale.io/v2/ltx-2.5/fast" \
|
|
| 76 |
--output "ltx2-5-fast.mp4"
|
| 77 |
```
|
| 78 |
|
| 79 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
-
Codex, ChatGPT, Claude, Cursor, Gemini CLI, and Antigravity can
|
| 82 |
|
| 83 |
-
|
|
|
|
|
|
|
| 84 |
|
| 85 |
---
|
| 86 |
|
| 87 |
-
## Links
|
| 88 |
|
| 89 |
- **Website**: [https://tensorscale.io](https://tensorscale.io)
|
| 90 |
-
- **
|
| 91 |
-
- **API
|
| 92 |
- **Pricing**: [https://tensorscale.io/pricing.html](https://tensorscale.io/pricing.html)
|
| 93 |
-
- **
|
| 94 |
-
- **Careers
|
|
|
|
| 9 |
|
| 10 |
<div align="center">
|
| 11 |
|
| 12 |
+
# TensorScale
|
| 13 |
|
| 14 |
+
### **Powering the future of generative worlds**
|
| 15 |
|
| 16 |
*The AI stack was built for language models. We’re rebuilding it for world models.*
|
| 17 |
|
| 18 |
[](https://tensorscale.io)
|
| 19 |
[](https://console.tensorscale.io)
|
| 20 |
+
[](https://tensorscale.io/docs.html)
|
| 21 |
[](https://tensorscale.io/pricing.html)
|
| 22 |
[](https://tensorscale.io/blog.html)
|
| 23 |
+
[](https://tensorscale.io/careers.html)
|
| 24 |
[](https://x.com/TensorScaleAI)
|
| 25 |
[](https://www.linkedin.com/company/tensorscale-ai)
|
| 26 |
|
|
|
|
| 28 |
|
| 29 |
---
|
| 30 |
|
| 31 |
+
## About Us
|
| 32 |
|
| 33 |
+
AI is entering a new era. Models are evolving from understanding and generating text to understanding and simulating the world itself—its geometry, physics, causality, dynamics, and interactions.
|
| 34 |
|
| 35 |
+
While the models have changed, the infrastructure has not. Today’s training and inference stack was designed for autoregressive language models. World models process and generate images, video, audio, actions, and 3D environments, demanding orders of magnitude more compute and fundamentally different execution patterns.
|
| 36 |
|
| 37 |
+
We are a team of experienced engineers and researchers with a track record of building world-class AI infrastructure at **Fireworks AI, Verkada, Snowflake, Meta, Google, Alibaba, Apple, and Microsoft**.
|
| 38 |
+
|
| 39 |
+
- **Generative media** — Enable real-time, interactive experiences at a fraction of the cost.
|
| 40 |
+
- **Robotics** — Generate robotic learning data at production scale.
|
| 41 |
|
| 42 |
---
|
| 43 |
|
| 44 |
+
## Performance — Unmatched speed and efficiency
|
| 45 |
|
| 46 |
+
**MiniMax H3 / 720p / 5s Video**:
|
| 47 |
+
- **10× faster** (`142 sec` baseline → **`13 sec`** on TensorScale)
|
| 48 |
+
- **Half the cost** (at **\$0.04 per sec**)
|
| 49 |
|
| 50 |
---
|
| 51 |
|
| 52 |
+
## Developer Experience — From API key to first generation in seconds
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
+
The TensorScale API ([`https://api.tensorscale.io`](https://tensorscale.io/docs.html)) serves hosted video and image generation models over HTTPS:
|
| 55 |
|
| 56 |
+
| Model | Model ID | Tasks | Synchronous | Streaming | Asynchronous |
|
| 57 |
+
| :--- | :--- | :--- | :---: | :---: | :---: |
|
| 58 |
+
| **MiniMax H3** | `MiniMax-H3` | Text, frames and references | Yes | Yes | Yes |
|
| 59 |
+
| **LTX-2.5 Fast** | `ltx2-5-fast` | Text and image to video | Yes | Yes | Yes |
|
| 60 |
+
| **Wan 3.0** | `wan-3.0` | Generation and editing | — | — | Yes |
|
| 61 |
+
| **MiniMax H3 Image** | `MiniMax-H3-Image` | Generation and editing | Yes | — | Yes |
|
| 62 |
+
| **Qwen Image 2.1** | `Qwen-Image-2.1` | Generation and editing | Yes | — | Yes |
|
| 63 |
+
| **MOSS Image** | `MOSS-Image-Base`, `MOSS-Image-SFT` | Text to image | Yes | — | Yes |
|
| 64 |
+
| **SenseNova U1.5** | `SenseNova-U1.5` | Generation and editing | Yes | — | Yes |
|
| 65 |
+
| **FLUX3 Image Klein** | `FLUX3-Image-Klein` | Generation and editing | Yes | — | Yes |
|
| 66 |
|
| 67 |
+
### `curl`
|
| 68 |
|
| 69 |
```bash
|
| 70 |
curl "https://api.tensorscale.io/v2/ltx-2.5/fast" \
|
|
|
|
| 83 |
--output "ltx2-5-fast.mp4"
|
| 84 |
```
|
| 85 |
|
| 86 |
+
### `Python`
|
| 87 |
+
|
| 88 |
+
```python
|
| 89 |
+
import os
|
| 90 |
+
from pathlib import Path
|
| 91 |
+
import requests
|
| 92 |
+
|
| 93 |
+
resp = requests.post(
|
| 94 |
+
"https://api.tensorscale.io/v2/ltx-2.5/fast",
|
| 95 |
+
json={
|
| 96 |
+
"prompt": "A cinematic aerial push toward a lighthouse as waves roll against the cliffs.",
|
| 97 |
+
"width": 1280,
|
| 98 |
+
"height": 768,
|
| 99 |
+
"num_frames": 121,
|
| 100 |
+
"frame_rate": 24,
|
| 101 |
+
"seed": 10,
|
| 102 |
+
},
|
| 103 |
+
headers={
|
| 104 |
+
"Authorization": f"Bearer {os.environ['TENSORSCALE_API_KEY']}",
|
| 105 |
+
"Accept": "video/mp4",
|
| 106 |
+
},
|
| 107 |
+
timeout=7200,
|
| 108 |
+
)
|
| 109 |
+
resp.raise_for_status()
|
| 110 |
+
Path("ltx2-5-fast.mp4").write_bytes(resp.content)
|
| 111 |
+
```
|
| 112 |
+
|
| 113 |
+
### Agent-native setup — Ask your agent to install TensorScale MCP
|
| 114 |
|
| 115 |
+
Codex, ChatGPT, Claude, Cursor, Gemini CLI, and Antigravity can follow one trusted setup skill to configure Streamable HTTP and OAuth—without putting API keys in chat:
|
| 116 |
|
| 117 |
+
```text
|
| 118 |
+
Read https://tensorscale.io/setup-mcp-skills.md and configure TensorScale MCP for me.
|
| 119 |
+
```
|
| 120 |
|
| 121 |
---
|
| 122 |
|
| 123 |
+
## Links
|
| 124 |
|
| 125 |
- **Website**: [https://tensorscale.io](https://tensorscale.io)
|
| 126 |
+
- **Console**: [https://console.tensorscale.io](https://console.tensorscale.io)
|
| 127 |
+
- **API Reference**: [https://tensorscale.io/docs.html](https://tensorscale.io/docs.html)
|
| 128 |
- **Pricing**: [https://tensorscale.io/pricing.html](https://tensorscale.io/pricing.html)
|
| 129 |
+
- **Blog**: [Rebuilding the AI Stack for the Era of World Models](https://tensorscale.io/blog-rebuilding-ai-stack-world-models.html)
|
| 130 |
+
- **Careers**: [https://tensorscale.io/careers.html](https://tensorscale.io/careers.html)
|