Adoption

Use Cases

Company

Features

Block-based workflow editor

Prompting, testing, versioning, multi-prompt, cost control

Resume line

Retrieval Augmented Generation (RAG)

Roles and permissions

Security, data transparency

Blocks

Web

Data Transformer

Decision Maker

Extract Maker (with HTML)

Key / Knowledge Manager

Web Creator

Hosting

SaaS Email, Integration

Cloud privacy managed

On premise

The ARI Platform

Solutions

ARI is made of two complementary layers: Flows, to build custom AI processes end-to-end, and Modules, to activate ready-to-play vertical features. Choose where to start.

Flows

Fully customizable AI processes end-to-end. For those who want to build.

→

Modules

Ready-to-play vertical features. For those who want to start right away.

→

On-premise available

LLM-agnostic

Zero vendor lock-in

Two layers, one platform

Build or activate. Always in control.

ARI FLOWS

Build AI processes on your terms. With full control.

ARI Flows is the heart of the platform: a visual environment for designing AI pipelines end-to-end. Choose your models, define your nodes, connect your data sources — ERP integrations, databases, APIs, documents. You can build any AI process that exists in your organization, with full visibility over every step. No vendor lock-in, no architectural compromise.

Two pillars of ARI Flows

Design and control

Every flow is a visual pipeline: choose AI models, connect nodes, define the logic. From prompt engineering to integration with your systems — every step configurable and transparent.

Test and maintain

Before going live, ARI verifies every flow. And keeps doing it automatically: change a model, update a data source, ARI alerts you immediately if something breaks — before any user notices.

→

Connect to APIs, databases, ERPs, CRMs — without building infrastructure from scratch.

HOW TO WORK WITH ARI

From problem to production, step by step.

ARI structures the AI development process into clear phases. Each phase has dedicated tools — so nothing is skipped and you iterate methodically.

Step 1 — Design

Define the objective and map the flow on the visual node editor. Describe what goes in, what should come out, and how you will measure success — before you even start building.

Step 2 — Develop

Configure inputs, set metrics, and build workflow variants. Connect nodes and building blocks, compare results, and refine each node until the flow behaves exactly as you want.

Step 3 — Validate

Choose the most suitable LLM model and launch automated test cycles. The Research Tree records every version: you can compare different approaches and roll back at any time.

Step 4 — Publish

When results satisfy you, publish with one click. ARI generates API documentation and integrates the flow into existing systems — SaaS, private cloud, or on-premise.

Step 5 — Monitor

Keep track of performance, LLM costs, and resource usage from a single real-time dashboard.

THE ARCHITECTURE

Flows, nodes, building blocks.

A flow is a complete business process — it defines what comes in, what should come out, and how. Each flow is made up of nodes and building blocks, at the same level. Nodes are customizable processing units (logic, integrations, AI, data). Building blocks are pre-configured, field-tested flows, ready to use — reusable across multiple flows with one click.

Input

Nodo

Logica · AI

Nodo

Dati · Integr.

Building

Block

Output

Nodo

Building Block

BUILDING BLOCKS

Use our ready-made blocks or create your own with flows.

Each building block solves a specific problem. Choose the ones you need, combine them in the workflow and reuse them in other processes without rewriting anything. Create your own Custom Block and reuse it as part of new flows whenever you want. Some of our Main Building Blocks:

ARI DB

Enables interaction with databases in natural language, supporting read and write operations on structured data.

Web Crawler

Extracts information from websites and returns relevant results based on a query.

Decision Maker — Basic

Makes decisions autonomously and makes the decision-making process transparent.

Decision Maker — Numeric Multimodal

Makes decisions autonomously and makes the decision-making process transparent, returning a numeric score.

Decision Maker — Binary Multimodal

Makes decisions autonomously and makes the decision-making process transparent, returning a boolean value.

Knowledge Base — Ingestion

Complete pipeline for document ingestion into RAG systems (chunking, indexing, etc.).

Knowledge Base — Normalize PDF

Normalizes PDFs to prepare them for ingestion (cleaning, structure, formatting).

Knowledge Base — Query

Performs advanced queries on RAG systems combining retrieval and AI. Native hybrid search: vector search (semantic) + BM25 (keyword) combined. Most RAG platforms force you to choose one; ARI runs both and merges the results.

Knowledge Base — Remove File

Removes documents and related data from RAG pipelines.

Custom Building Block

Any flow developed in ARI can be turned into a block and used as part of a new flow.

BUILT ON ARI

Built on ARI. And we prove it.

Built into ARI is a conversational AI assistant: it helps you build workflows, explains what a node does, suggests optimizations, guides you through debugging.

That assistant is an ARI workflow in execution. Same node editor. Same workflow runtime. Same multi-agent streaming. Same Model Context Protocol for calling external tools. Same persistent conversation history.

There's no special infrastructure, no second product: anything our assistant does, yours can do too. Want an AI assistant integrated into your enterprise app? Build it like we did: a conversational workflow, a node, MCP, done.

If ARI is solid enough to run our AI in production, it's solid enough for yours.

MODULES

Four vertical products. Zero development.

Modules are AI products already configured and ready to use. Each one solves a specific problem — querying databases, searching internal documents, creating an internal AI assistant, extracting data from any file. You can activate them without accessing the rest of the platform, through a dedicated interface designed for the people who use them every day.

How they differ from ARI Flows

ARI Flows

Build custom AI workflows, node by node. For those who want to create tailored processes that don't exist yet — with full control over every step.

ARI Modules

Use a product already built and optimized for a specific use case. For those who want to solve a precise problem starting today, with no development required.

↔

The two approaches are combinable: you can start with a Module and then extend with custom Flows.

European low-code GenAI orchestration platform. Built for control, transparency and trust.

Black Bytes SA — Via Pretorio 13A, 6900 Lugano, Switzerland

Get in touch

Cookies Policy

Cookies and tracking

We use technical cookies and — with your consent — analytics and marketing cookies to measure traffic and improve your experience. You can accept, refuse or change your choice at any time.

More info in the Cookies Policy →

The ARI Platform

Solutions

ARI is made of two complementary layers: Flows, to build custom AI processes end-to-end, and Modules, to activate ready-to-play vertical features. Choose where to start.

Flows

Fully customizable AI processes end-to-end. For those who want to build.

→

Modules

Ready-to-play vertical features. For those who want to start right away.

→

On-premise available

LLM-agnostic

Zero vendor lock-in

Two layers, one platform

Build or activate. Always in control.

ARI FLOWS

Build AI processes on your terms. With full control.

Two pillars of ARI Flows

Design and control

Every flow is a visual pipeline: choose AI models, connect nodes, define the logic. From prompt engineering to integration with your systems — every step configurable and transparent.

Test and maintain

Before going live, ARI verifies every flow. And keeps doing it automatically: change a model, update a data source, ARI alerts you immediately if something breaks — before any user notices.

→

Connect to APIs, databases, ERPs, CRMs — without building infrastructure from scratch.

HOW TO WORK WITH ARI

From problem to production, step by step.

ARI structures the AI development process into clear phases. Each phase has dedicated tools — so nothing is skipped and you iterate methodically.

Step 1 — Design

Define the objective and map the flow on the visual node editor. Describe what goes in, what should come out, and how you will measure success — before you even start building.

Step 2 — Develop

Configure inputs, set metrics, and build workflow variants. Connect nodes and building blocks, compare results, and refine each node until the flow behaves exactly as you want.

Step 3 — Validate

Choose the most suitable LLM model and launch automated test cycles. The Research Tree records every version: you can compare different approaches and roll back at any time.

Step 4 — Publish

When results satisfy you, publish with one click. ARI generates API documentation and integrates the flow into existing systems — SaaS, private cloud, or on-premise.

Step 5 — Monitor

Keep track of performance, LLM costs, and resource usage from a single real-time dashboard.

THE ARCHITECTURE

Flows, nodes, building blocks.

Input

Nodo

Logica · AI

Nodo

Dati · Integr.

Building

Block

Output

Nodo

Building Block

BUILDING BLOCKS

Use our ready-made blocks or create your own with flows.

ARI DB

Enables interaction with databases in natural language, supporting read and write operations on structured data.

Web Crawler

Extracts information from websites and returns relevant results based on a query.

Decision Maker — Basic

Makes decisions autonomously and makes the decision-making process transparent.

Decision Maker — Numeric Multimodal

Makes decisions autonomously and makes the decision-making process transparent, returning a numeric score.

Decision Maker — Binary Multimodal

Makes decisions autonomously and makes the decision-making process transparent, returning a boolean value.

Knowledge Base — Ingestion

Complete pipeline for document ingestion into RAG systems (chunking, indexing, etc.).

Knowledge Base — Normalize PDF

Normalizes PDFs to prepare them for ingestion (cleaning, structure, formatting).

Knowledge Base — Query

Knowledge Base — Remove File

Removes documents and related data from RAG pipelines.

Custom Building Block

Any flow developed in ARI can be turned into a block and used as part of a new flow.

If ARI is solid enough to run our AI in production, it's solid enough for yours.

MODULES

Four vertical products. Zero development.

How they differ from ARI Flows

ARI Flows

Build custom AI workflows, node by node. For those who want to create tailored processes that don't exist yet — with full control over every step.

ARI Modules

Use a product already built and optimized for a specific use case. For those who want to solve a precise problem starting today, with no development required.

↔

The two approaches are combinable: you can start with a Module and then extend with custom Flows.