Modselect

Modselect is No Code AI Solution Platform. Visit www.modselect.com for more information.

Modselect simplifies the usage of AI models, providing a convenient way to interact with models and implement AI solutions into your website or application.

UPDATES:

v0.0.17 - Stable Diffusion UI has been added.

v0.0.14 - Support for Google Gemini API has been added, using the @google/generative-ai library. This is not using the newer Vertex AI API, but instead the previous library version. Support for Google's Vertex AI API will be in a future version of modselect.

v0.0.13 - you can now use your Hugging Face and OpenAI keys with yous custom solution. Visit the Solution Builder at www.modselect.com and assign your own API key to your solution deployment.

Usage

Installation

npm install modselect

pnpm install modselect

yarn add modselect

Import

import { Modselect } from "modselect"

Custom Solution Implementation

In order to use the Modselect library Solution tag, you must first create a Solution and Deployment in the WebUI. Use the mID and API_KEY from your saved deployment and place this tag in your website/app. Optionally, a debug tag is available which provides a simple popover for viewing Solution information.

<Modselect mID={mID} API_KEY={API_KEY} />
<Modselect mID={mID} API_KEY={API_KEY} debug />

Stable Diffusion

Now supporting a Stable Diffusion React UI, clone of AUTOMATIC1111. Runs against the normal route in the stable diffusion API: /sdapi/v1/img2img or /sdapi/v1/txt2img. Includes a URL text field for configuring your local server.

import { StableDiffusionUI } from "modselect"
<StableDiffusionUI  />

Development Mode

The library also contains a developmet mode providing easy to use UI for Hugging Face and Gemini.

import { HuggingFace, Gemini } from "modselect"

Hugging Face

<HuggingFace 
    type="image-to-text"
    API_KEY={HF_KEY}
    model="Salesforce/blip-image-captioning-large" />

<HuggingFace 
    type="text-to-text"
    API_KEY={HF_KEY}
    model="openai-community/gpt2" />

<HuggingFace 
    type="text-to-audio"
    API_KEY={HF_KEY}
    model="facebook/mms-tts-eng" />

<HuggingFace 
    type="text-to-image"
    API_KEY={HF_KEY}
    model="cagliostrolab/animagine-xl-3.0" />

Google/Gemini

Google/Gemini development mode is currently not supported given the recent change to Vertex AI

<Gemini 
    type="image-to-text"
    API_KEY={GOOGLE_AI_KEY}
    model="gemini-pro-vision"/>

<Gemini 
    type="text-to-text"
    API_KEY={GOOGLE_AI_KEY}
    model="gemini-pro"/>