EnergeticAI

EnergeticAI is a powerful open-source library designed to make it easy for developers to use AI models directly within their Node.js applications. It serves as a highly optimized distribution of TensorFlow.js, tailored specifically for the constraints of serverless environments like AWS Lambda or Vercel Functions. The primary goal of EnergeticAI is to overcome the common challenges of using machine learning models in serverless architectures, such as large module sizes and slow cold-start times, which can lead to high latency and poor user experience.

By providing a condensed, high-performance replacement for standard TensorFlow.js, EnergeticAI enables developers to build sophisticated AI-powered features with minimal overhead. It comes with pre-trained models for common Natural Language Processing (NLP) tasks, allowing for rapid implementation without the need for extensive machine learning expertise or reliance on costly and proprietary third-party APIs. This self-hosted approach ensures full control over data, enhances privacy, and eliminates vendor lock-in.

How to use EnergeticAI

Getting started with EnergeticAI is straightforward and designed for a seamless developer experience. It can be installed directly from NPM into any Node.js project (version 18+).

1. Installation: Install the core package along with the desired model libraries. For example, to use text embeddings, you would run:

npm install @energetic-ai/core @energetic-ai/embeddings @energetic-ai/model-embeddings-en

2. Initialize the Model: In your application code, import and initialize the model. To improve cold-start performance, it's recommended to use the pre-bundled model source.

import { initModel } from "@energetic-ai/embeddings";
import { modelSource } from "@energetic-ai/model-embeddings-en";
const model = await initModel(modelSource);

3. Generate Embeddings: Use the initialized model to convert text into vector embeddings. You can process a single string or an array of strings.

const embeddings = await model.embed(["Fruit is healthy", "Fruit is delicious"]);

4. Perform Classification: For text classification, initialize the classifier with a few examples (few-shot learning) and then classify new text.

import { initClassifier } from "@energetic-ai/classifiers";
const classifier = await initClassifier([["I had a great day", "Positive"], ["I am frustrated", "Negative"]]);
const result = await classifier.classify("The weather is so nice today");

5. Integrate with Vector Databases: The generated embeddings can be stored and queried in vector databases like Postgres (with pgvector), Milvus, Chroma, or Pinecone to build powerful semantic search or recommendation systems.

Core Features of EnergeticAI

• Optimized for Serverless: Features an extremely small module size (~3 MB) and incredibly fast cold-start inference (~50 ms), making it up to 67x faster than standard TensorFlow.js in serverless environments.
• Text Embeddings: Utilizes a lightweight version of Google's Universal Sentence Encoder to convert English text into 512-dimensional vectors, capturing semantic meaning for similarity comparisons, search, and recommendations.
• Few-Shot Text Classification: Allows for the classification of text into custom categories with just a few training examples per category, eliminating the need for large datasets and model retraining.
• Easy Integration: A simple NPM package that acts as a drop-in replacement for `@tensorflow/tfjs-core`, enabling incremental adoption in existing projects.
• Business-Friendly Licensing: Licensed under Apache 2.0, making it free to use for commercial purposes without the restrictions of proprietary APIs.
• Self-Hosted and Private: Models run within your own infrastructure, ensuring that sensitive user data is never sent to third-party services.

Use Cases for EnergeticAI

EnergeticAI is ideal for developers looking to add intelligence to their applications without external dependencies. Common use cases include:

• Semantic Search: Build search engines that understand the meaning and context of queries, not just keywords.
• Recommendation Engines: Suggest relevant products, articles, or content to users based on semantic similarity.
• Automated Ticket Routing: Classify customer support tickets into categories like "Billing," "Technical," or "Sales" to streamline workflows.
• Content Moderation: Automatically categorize user-generated content to flag inappropriate or spammy submissions.
• Sentiment Analysis: Analyze customer feedback or social media posts to gauge sentiment (e.g., Positive/Negative).

Advantages of EnergeticAI

Compared to alternatives like stock TensorFlow.js or proprietary APIs (OpenAI, Cohere), EnergeticAI offers significant advantages:

• Superior Performance: Specifically engineered to solve the cold-start and package size problems inherent in serverless functions.
• Cost Efficiency: As an open-source library, it eliminates recurring API costs, making it highly economical for projects at any scale.
• Data Privacy and Control: Keeps all data processing in-house, which is crucial for applications handling sensitive information.
• No Vendor Lock-in: Frees your business from dependency on a single provider's pricing, terms of service, and platform idiosyncrasies.
• Simplicity and Ease of Use: Offers a minimal API surface and pre-trained models, lowering the barrier to entry for implementing AI features.

Pricing and Plans

EnergeticAI is completely free and open-source, distributed under the permissive Apache 2.0 license. There are no paid plans, subscription fees, or usage-based costs associated with the library itself. Developers can use it in personal and commercial projects without any financial obligation.