Blame for README.md

---
pipeline_tag: text-generation
language:
  - en
  - it
tags:
  - chronos-dx
  - flushnet
  - large-language-model
  - text-generation
  - instruction-following
  - local-ai
  - private-ai
  - secure-ai
  - cloud-ai
  - scalable-ai
  - browser-ai
  - llm-fine-tuning
  - model-training
  - enterprise-ai
pretty_name: Chronos DX
---

# Chronos DX

**Chronos DX** is the Flushnet AI model family for private, customizable and production-oriented generative AI.

Chronos DX is designed for two complementary deployment modes:

- **Local / Private / Secure AI** — for customer-controlled environments, local execution, privacy-focused workloads and reduced cloud dependency.
- **Cloud / High-Scalability AI** — for managed APIs, high concurrency, larger workloads, autoscaling and production services.

The same Chronos DX family can therefore be used from private local inference through to highly scalable cloud deployments.

## Try Chronos DX

### Local Secure Demo

https://www.flushnet.net/chronos-secure/secure-webchat/

The current browser-based preview demonstrates local/private Chronos DX execution in a supported browser environment.

### Chronos DX Model Training & Customization

https://www.flushnet.net/ai/chronos-dx/

Project requests:

https://www.flushnet.net/ai/chronos-dx/#contact

## Chronos-DX demos: two flavours

Chronos-DX demos are available in **two flavours**:

1. **[Local Private Secure Browser Demo](https://www.flushnet.net/chronos-secure/secure-webchat/)** — a compressed, browser-based version designed for local and private use.
2. **[Dense 32B GGUF https://huggingface.co/FlushnetAI/Chronos-DX-Dense-32B-GGUF](https://huggingface.co/FlushnetAI/Chronos-DX-Dense-32B-GGUF)** — the full dense model distributed as a downloadable GGUF for local inference, with **32.085B total parameters (~32B active)**.

Explore [Chronos-DX on Flushnet](https://www.flushnet.net/ai/chronos-dx/) and the [Chronos-DX Ollama listing](https://ollama.com/marcheseluciano/Chronos-DX). The Ollama listing is a project page; local model weights are currently provided by the GGUF repository above.

## Knowledge specialization and custom training

Chronos-DX can be **highly customized on request** for specific fields of knowledge, organization-specific data, workflows and customer requirements. Available services include custom training, fine-tuning, domain specialization, knowledge-base (RAG) integration and targeted model improvements.

**[Request training or model customization](https://www.flushnet.net/chronos-dx-model-request/)**

## Model Family

Chronos DX supports model engineering across the following model-size classes.

| Model class | Availability |
|---|---|
| Chronos DX 8B | Current family reference; local secure preview available |
| Chronos DX 12B | Custom training / model build available |
| Chronos DX 20B | Custom training / model build available |
| Chronos DX 32B | [Dense 32B GGUF download](https://huggingface.co/FlushnetAI/Chronos-DX-Dense-32B-GGUF) and custom training |
| Chronos DX 40B | Custom training / model build available |
| Chronos DX 70B | Custom training / model build available |
| Chronos DX 80B | Custom training / model build available |

## Capabilities

Depending on the selected Chronos DX model, customization and deployment target, capabilities can include:

- instruction following
- general text generation
- coding and software-engineering assistance
- infrastructure and cloud-oriented tasks
- enterprise and domain specialization
- structured outputs
- multilingual workloads
- private/local inference
- scalable cloud inference
- customer-specific model adaptation
- model retraining and updates
- LoRA / QLoRA fine-tuning
- supervised fine-tuning
- model evaluation and regression testing
- GGUF and supported quantized artifacts
- local deployment support
- cloud deployment support
- private enterprise deployment

## Deployment Options

### Local / Private / Secure

Chronos DX can be deployed in customer-controlled environments for workloads where privacy, local processing, data control or reduced cloud dependency are important.

Possible deployment targets include:

- browser-local AI
- private workstations
- local servers
- edge systems
- customer-controlled infrastructure
- private enterprise environments

### Cloud / Highly Scalable

Chronos DX can also be deployed on cloud infrastructure for production workloads requiring scale and centralized management.

Possible deployment targets include:

- managed model APIs
- private cloud endpoints
- autoscaled inference services
- dedicated cloud instances
- multi-instance deployments
- high-concurrency applications
- enterprise cloud environments

### Hybrid

Chronos DX can also support hybrid architectures combining local/private inference with cloud capacity.

This allows selected workloads to remain local while larger, burst or high-concurrency workloads can use cloud infrastructure.

## Training and Customization

Chronos DX can be trained, fine-tuned, specialized, updated and prepared for customer-specific requirements.

Available services include:

- automated fine-tuning
- managed fine-tuning
- domain specialization
- supervised fine-tuning
- LoRA / QLoRA
- model updates and retraining
- dataset preparation
- training dataset engineering
- synthetic dataset generation
- model evaluation
- benchmark testing
- model merge
- GGUF and quantization
- production validation
- local deployment preparation
- cloud deployment preparation
- enterprise customization

Full information and project requests:

https://www.flushnet.net/ai/chronos-dx/

## Intended Uses

Chronos DX can be adapted for applications such as:

- coding assistants
- software engineering
- infrastructure and cloud automation
- technical support
- enterprise knowledge assistants
- domain-specific assistants
- customer-service workflows
- structured information extraction
- private/local AI
- secure enterprise AI
- multilingual assistants
- scalable cloud AI services
- hybrid local/cloud AI systems

## Versioning

Chronos DX model-specific releases use explicit version/build identifiers so model variants and deployment artifacts can be tracked independently.

The Chronos DX family page provides the general overview, while specific editions can receive dedicated repositories as they become available.

## About Flushnet AI

Chronos DX is developed by **Flushnet AI**.

Chronos DX information, model engineering and training requests:

https://www.flushnet.net/ai/chronos-dx/

Local secure demo:

https://www.flushnet.net/chronos-secure/secure-webchat/