# Dynamixs AI

Dies ist das interne Super Office von Dynamixs.ai

Wir verwenden OIDC / Keyclaok für die Anmeldung

This repository contains the source code as also build scripts to setup an Dynamixs AI instance.

**Topics**

- [Imixs Worklflow Engine - BPMN 2.0](./doc/IMIXS_WORKFLOW_EN.md)
- [Retrieval Augmented Generation (RAG)](./doc/RAG.md)
- [A2A Agent Integration](./doc/A2A.md)
- [Knowledge DB (Apache Cassandra)](./doc/RAG_CASSANDRA.md)

## Build & Run

The dev environment is defined by the 'docker compose' file `docker-compose-dev.yaml`

The Environment contains the following components:

- Imixs-Office-Workflow - application instance based on Wildfly 29.x
- PostgreSQL Server - postgres:9.6.1 - local database
- Apache Tika - apache/tika:3.2.0.0-full for OCR
- Apache Cassandra - cassandra:5.0.4 local vector database
- Collabora - collabora/code:25.04.6.2.1 for local office documents
- Imixs-Admin - imixs/imixs-admin Imixs Admin Client
- Prometheus - prom/prometheus:latest for logging metrics
- Grafana - grafana/grafana:10.4.0 - for metric monitoring

### AI / LLM

The system expects a LLM Server endpoint based on Open AI API. The default LLM server endpoint is defined by the environment variables `LLM_SERVICE_ENDPOINT` and `LLM_API_KEY`. You can define both variables in the file `./docker/.env`

```
LLM_API_ENDPOINT=https://my.llama.cpp.foo.com/
LLM_API_KEY=your-api-key-here
```

### Build / Setup

To build and run the project you can use the build script `devi.sh` and run:

    $ ./devi setup

This wild build the project with 'maven' and also build the local docker image. The project will be started automatically.

You can also start the project with

    $ ./devi start

To access the application from our Web browser

    http://localhost:8080

(Default user `admin' - default passwort `adminadmin`)