Wool Ball 🧶 Wool Ball

Wool Ball is an open-source AI platform powered by a distributed network of browsers. It offers access to AI models and tasks at a fraction of typical prices, leveraging a collaborative compute model where contributors help process tasks by keeping their browser sessions active.

---

What it is

• An open-source AI ecosystem that runs via a distributed browser network.
• Provides 10 AI task types at about 10% of the price of competing services.
• Aims to reduce costs for users and reward contributors who keep their browsers open to process tasks.

---

AI Task Types and Pricing

• llava-onevision-qwen2-0.5b: Text to Text, Image+Text to Text
• SmolLM2-135M-Instruct: Text Generation (HuggingFaceTB)
• SmolLM2-360M-Instruct: Text Generation (HuggingFaceTB)
• Qwen2.5-0.5B-Instruct: Text Generation (Mozilla)
• Qwen2.5-Coder-0.5B-Instruct: Text Generation (Qwen)
• mms-tts: Text to Speech (PT-BR, ES, EN) costs $1.50 per 1M chars
• whisper-large-v3-turbo: Speech to Text (timestamped, 99 languages)
• nllb-200-distilled-600M: Translation (196 languages)
• xLM-Roberta-large-XNLI: Zero Shot Classification
• facial_emotions_image_detection: Image Classification (facial emotions)
• distilbart-cnn-6-6-optimised: Summarization

Pricing sample (in/out):

• Text in: $0.016 / 4M chars (example model) | Text out: $0.024 / 4M chars
• Text to Speech: $1.50 / 1M chars
• Speech to Text: per minute rate provided

---

How it Works

• Models are hosted in a distributed network of participant browsers.
• Users submit tasks (text generation, translation, TTS/STT, classification, summarization, etc.).
• Contributors keep their browsers open to process tasks and are rewarded for the processed workload.

---

Getting Started

1. Visit Wool Ball and create an account if required.
2. Choose an AI task type and submit your input data.
3. Retrieve the generated output and, if desired, contribute by keeping your browser active to earn rewards.

---

Usage Examples

• Text Generation, Translation, and Summarization tasks with multiple models
• Speech-to-Text and Text-to-Speech workflows
• Image Classification and Facial Emotion Analysis

---

Community & Participation

• Become a contributor by keeping your browser open and participating in the distributed compute network.
• Earn rewards for processing AI tasks.
• Engage with the community via Discord and reach out via [email protected].

---

Core Features

• Free to access core AI task types via a distributed browser network
• 10 AI task types at ~10% of competitor pricing
• Wide range of tasks: text generation, translation, TTS, STT, image classification, zero-shot classification, summarization, and more
• Open-source model catalog with ongoing contributions from the community
• Rewards for contributors who keep their browsers processing tasks
• Community support and direct contact options (Discord, email)