Spaces:
Running
Running
Add organization card
Browse files
README.md
CHANGED
|
@@ -1,10 +1,44 @@
|
|
| 1 |
---
|
| 2 |
title: README
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
title: README
|
| 3 |
+
emoji: ⚡
|
| 4 |
+
colorFrom: gray
|
| 5 |
+
colorTo: indigo
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
| 8 |
---
|
| 9 |
|
| 10 |
+
# Everypixel
|
| 11 |
+
|
| 12 |
+
Everypixel builds computer vision and generative media tools. We also run blind benchmarks of AI models.
|
| 13 |
+
|
| 14 |
+
## Everypixel Labs API
|
| 15 |
+
|
| 16 |
+
Pay-as-you-go endpoints with no subscription. You get 500 free requests, no card required.
|
| 17 |
+
|
| 18 |
+
| Endpoint | What it does | Price |
|
| 19 |
+
|---|---|---|
|
| 20 |
+
| `/image_keywording` | Image tagging in 7 languages, plus color recognition | $0.6 / 1,000 requests |
|
| 21 |
+
| `/image_captioning` | Detailed alt text from an image | $1.2 / 1,000 requests |
|
| 22 |
+
| `/photo_quality` | Scores and ranks images by quality | $0.6 / 1,000 requests |
|
| 23 |
+
| `/age_recognition` | Estimates a person's age from a face | $0.6 / 1,000 requests |
|
| 24 |
+
| `/text_to_speech` | Converts text to natural speech | $0.02 / 60 sec |
|
| 25 |
+
| `/lipsync` | Syncs a voiceover with video | $1 / 60 sec |
|
| 26 |
+
| `/face_swap` | Swaps faces in images and video | $10 / 1,000 requests |
|
| 27 |
+
|
| 28 |
+
[API overview](https://labs.everypixel.com/api) · [Docs](https://labs.everypixel.com/docs)
|
| 29 |
+
|
| 30 |
+
## Everypixel Research
|
| 31 |
+
|
| 32 |
+
Blind pairwise evaluations of image generation and upscaling models, judged by working professionals. Recent tests:
|
| 33 |
+
|
| 34 |
+
- **GPT Image 2 vs real photography:** GPT Image 2, Qwen Image 2, and FLUX.2 Klein 9B against real photos. 48 scenes, 13 professional judges.
|
| 35 |
+
- **SeedVR2 vs Magnific:** 264 image pairs, 13 evaluators, 792 judgments. SeedVR2 led on detail recovery (86%) and naturalness (82%).
|
| 36 |
+
- **SeedVR2 vs Topaz Wonder 3:** 264 pairs. SeedVR2 was stronger on edges, Topaz on faces.
|
| 37 |
+
|
| 38 |
+
[All benchmarks](https://research.everypixel.com/)
|
| 39 |
+
|
| 40 |
+
## Everypixel Workroom
|
| 41 |
+
|
| 42 |
+
A web workspace for image, video, voice, and lipsync generation. It gives access to GPT Image 2, FLUX.2, Seedream 4.0, Kling 3.0, Seedance 2.0, Runway Gen 4.5, WAN 2.7, and other models. The free plan includes 200 credits.
|
| 43 |
+
|
| 44 |
+
[Try Workroom](https://workroom.everypixel.com/)
|