Multimodal

LLM-powered apps

10x faster

NiddamHub is an open-source toolkit to help developpers build LLM-powered applications for any project.

Start for free

Connect your LLMs providers

AI agents

Create multimodal AI agents that can interact with your data & APIs and generate voice responses in real-time

Structured data generation

Generate validated JSON based on any schema from multiple data sources including text, images, audio, and URLs.

Long-term memory

No more need to manage conversations history, we did it for you with managed Serverless encrypted memory stores

Knowledge stores

Expand your AI agents knowledge by uploading files or websites to Serverless knowledge stores served edge-close to users

Functions

Custom function from your code

Google

Search google for resutls

Your API

Connect any API you want

APIs & Tools

Equip your AI agents with APIs and tools they can use to fetch data or perform actions based on context

Voice interaction

Accept voice inputs, process multiple audio files in parallel, and stream voice responses in under 300ms

Clean data & store embeddings

Served on the edge to your agents

Knowledge Stores

Upload files, PDFs, or websites to knowledge stores to expand your AI agents knowledge

Speed

Security

Accuracy

Fast & Reliable

NiddamHub is optimized for performance and real-time interactive applications, with built-in streaming, memory encryption, LLM output validation, and smart errors recovery

Scale globally

Effortlessly scale on your own rules. Both knowledge & memory stores are scalable and served edge-close to your users from 26 regions across the globe

Vision

Just simply use a LLM that supports vision and send your images as inputs without any additional setup

Your message...

You

Is this a paid platform?

AI Assistant

Based on the knowledge we have

We have a free plan so you can use its core features on your servers totally for free, but it provides paid options with more features like long-term memory and knowledge stores

Playground

AI agents can be used to build automation apps, AI conversational bots (text & voice), or for data extraction tasks.

Validated JSON data

Data Extraction

Extract data from multiple data sources and generate a validated JSON object based on any data schema

Simple Process

00. Connect your provider

Connect the LLM provider that will power your AI agents safely from your servers and never share them with our servers

Create AI agent

01. Create AI bot/agent

AI agents can be used to build automation apps, AI conversational bots (text & voice), or for data extraction tasks.

import { Scoopika, Agent } from '@scoopika/scoopika'

const scoopika = new Scoopika()
const agent = new Agent('ID', scoopika)

agent.run({
    inputs: {
        message: "Schedule a call..",
        audio: [],
        images: [],
        urls: []
    },
    hooks: {
        onToken: (token) => {...},
        onToolCall: (call) => {...},
    }
})

02. Build

Use the agent in your app or deploy it as an API endpoint in seconds with your favorite web framework.

Why Niddam-Hub?

Built for developers

Built for engineers, developers, dreamers, thinkers and doers.

Ease of use

Just send inputs to your AI agent, much easier than sending a fetch request

LLM Output validation

Built-in validation for JSON object generation with retries and error handling

Real-time streaming

Stream text & voice responses as they're being generated

Affordable

Introduces $0 to your LLMs costs. Only pay for extra useful features

Host anywhere

Host your own niddam-hub endpoint on your favorite cloud. memory & knowledge stores are on us ;)

Managed memory

No more need to manage conversations with AI and store messages yourself, we got you covered!

Open source

Proudly open source for you and every builder to develop and use with freedom

Use cases

You can build any AI application with the OpenAI SDK, now 10 times faster, more reliable, with multimodal inputs.

Custom AI assistants for your website

Multimodal AI-powered search engine

Rabbit R1 (but the real software)

No code

Clean code

Built from the ground to be easily integrated in web applications, with simple APIs, full type-safety, and easy error handling

Works with any web framework, React, and NextJS

Streaming made easy with hooks

Deploy a niddam-hub endpoint with 3 lines of code

Github

All parts of Niddam-Hub that have been developed as open source

Pricing

HOBBY

For developers building their impressive LLM-powered side projects & apps

Free

Unlimited AI agents runs

Connect your LLMs providers

Add tools to agents

Real-time streaming

LLM output validation

Full type-safety

Smart error recovery

PRO

For developers & small teams that need voice responses, long-term memory, or agents with custom knowledge

$25

/month

Everything in the free plan, plus:

Serverless managed long-term memory

Expand agents knowledge from files & websites

Real-time voice responses

Higher rate limits

Faster audio files processing

SCALE

For bigger teams that need higher rate limits, requests, volumes, and memory storage

$70

/month

Everything in the Basic plan, plus:

Higher memory storage & requests

Higher knowledge monthly requests

More voice responses characters

Even higher rate limits (almost none)

Direct support and assistant

Special Offer

We'll build your full AI feature in a week, you name it, we build it!

You get 3 months of the scale plan, plus:

Build any AI feature you want with Niddam-Hub

We handle both server-side & client-side development

UI/UX is on us, based on your product design

Need full app? We got you too!

LIMITED DEAL

$750

Limited to only 10 clients, get your spot now!

Learn more about pricing

Let's start building right now! niddam-hub introduces $0 extra costs to your LLM provider's costs. Only pay for optional features that deserve paying for

Start building