> For the complete documentation index, see [llms.txt](https://alquimia.gitbook.io/alquimia-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://alquimia.gitbook.io/alquimia-docs/runtime.md). # Operational Handler [Community](#community) | [Key features](#key-features) | [Use cases](#use-cases) | [Getting started](#getting-started) | [Workflow](#workflow) | [Examples](#examples) | [Contributing](#contributing) | [Alquimia AI](https://www.alquimia.ai/)
## Community We are building **enterprise-ready AI agents** with a focus on **transparency, consistency, and flexibility**. Contributions from the community help ensure reliability and innovation. ### Get Involved * **GitHub Discussions** – [Join here](https://github.com/Alquimia-ai/documentation/blob/main/discussions/README.md) * **Slack** – [Community chat](https://slack.com/alquimiaai) Your contributions help improve AI automation for everyone. ## Introduction **Alquimia Operational Handler** is an advanced, event-driven platform designed to manage multi-agent **LLM (Large Language Model)** solutions in **containerized environments**. Built on **Knative**, it provides seamless orchestration of LLMs, intelligent memory management, context-aware prompting, and complex tool execution. Designed for **Openshift** and **Kubernetes**, the platform offers **lightweight deployment, high scalability, and native integrations** with modern AI ecosystems, including **Openshift AI** and **LangChain**. It supports a diverse range of **LLM providers**, **vector stores**, and **retrieval-augmented generation (RAG) strategies**, making it an ideal solution for enterprises and developers building AI-powered applications. *** ## Key Features ### 🧬 **Event-Driven & Serverless** * Built on **Knative** for automatic scaling and serverless execution. * Fully **asynchronous** to ensure optimal performance and responsiveness. ### :cloud: **Seamless Cloud-Native Integration** * Works **natively** on **Openshift** and **Kubernetes**. * Supports **Openshift AI** for direct access to deployed models. * **LangChain-compatible**, enabling powerful agent-driven workflows. ### 🦜 **Flexible Multi-LLM Support** * Works with major LLM providers, including: * **OpenAI** * **Mistral** * **DeepSeek** * **LLama** ### :book: **Advanced RAG & Vector Store Integration** * Supports **retrieval-augmented generation (RAG)** for enhanced AI reasoning. * Compatible with vector stores like: * **Qdrant** * **Chroma** * **ElasticSearch** ### :envelope\_with\_arrow: **Omnichannel AI Integration** * Use **custom connectors** or **community Kamelets (Camel K)** for seamless omnichannel support. * Automate AI-powered workflows across multiple communication channels. ### 🥷 **Versatile Tool Execution** * Supports **server-side**, **client-side**, and **hybrid** tool execution. * Context-aware execution strategies to optimize performance. ### :rocket: **Lightweight & Production-Ready** * Minimal boilerplate, enabling rapid development and deployment. * Enterprise-ready with **scalability, reliability, and observability**. *** ## Why Choose Alquimia Operational Handler? ✅ **Scalability** – Effortlessly scale AI workflows with Knative.\ ✅ **Flexibility** – Works with multiple LLM providers, vector stores, and orchestration frameworks.\ ✅ **Performance** – Asynchronous, event-driven execution optimizes efficiency.\ ✅ **Integration** – Native compatibility with **LangChain**, **Openshift AI**, and containerized environments.\ ✅ **Serverless Superpowers** – Automatically scale workloads, reducing operational costs. *** ## Use Cases 1. **Multi-Agent AI Orchestration** – Manage and coordinate complex LLM-driven workflows. 2. **Enterprise-Scale Document Retrieval** – Implement **RAG** for intelligent search and knowledge retrieval. 3. **Omnichannel AI Automation** – Deploy AI-powered solutions across multiple communication channels. 4. **Hybrid Tool Execution** – Dynamically execute AI tools across client, server, or hybrid environments. *** ## Getting Started ### **Prerequisites** * A running **Openshift** or **Kubernetes** cluster. * **Openshift Serverless (Knative)** runtime installed. * **Openshift Service Mesh (Istio)** for networking. * **AMQ Streams (Strimzi)** for event-driven messaging. * A Redis instance for memory and cache management * A Couchdb instance to manage agent configurations * Optional: **Vector store** (e.g., Qdrant, Chroma, or ElasticSearch) for RAG capabilities. ### **Installation** #### Install on Openshift Ensure you have: * **OC client** installed. * **Knative support for Kafka** via **Strimzi (AMQ Streams)**. Then, deploy the platform: ```bash oc apply -f serving/base.yaml oc apply -f eventing/base.yaml ``` Now you are ready to deploy your first agent ## Workflow The proposed architecture is intented to be a common framework for agents. You can change it to adapt your needs. Recommend set up: ```mermaid sequenceDiagram participant CC as Client participant AH as Hermes (Entrypoint) participant IB as Inbound Broker participant NB as Normalized Broker participant CB as Classified Broker participant AL as Alquimia Leviathan (Execution steps) participant OB as Outbound Broker participant SC as Slack Connector CC ->> AH: Client sends query via Slack AH ->> IB: Trigger agent inference IB ->> AL: Trigger normalization Sequence AL ->> AL: Get memory for current session AL ->> NB: Pass normalized event NB ->> AL: Trigger agent custom sequence AL ->> AL: Executes classification models or LLMs with different roles AL ->> CB: Pass classified event CB ->> AL: Trigger Empathy Sequence AL ->> AL: Waits for other events to complete (tool execution for example) AL ->> AL: Select best expert profile according to context AL ->> AL: Invokes final LLM with selected profile AL ->> OB: Pass outbound event OB ->> AL: Memory persistance OB ->> SC: Send answer via Slack ``` ## Examples See our list of full working examples [here](https://github.com/Alquimia-ai/documentation/blob/main/examples/README.md) ## Integrations You can find our custom channel integrations [here](https://github.com/Alquimia-ai/documentation/blob/main/runtime/integrations/README.md) or deploy [community Kamelets](https://camel.apache.org/camel-kamelets/4.8.x/) ## Server tools Find all server tools available [here](https://github.com/Alquimia-ai/documentation/blob/main/runtime/tools/README.md) and see how easy is to create your own bundle. ## Local development For more information on how to set your local develop environment see docs [here](https://github.com/Alquimia-ai/documentation/blob/main/runtime/%27docs/local-development.md%27) ## CLI Usage Alquimia Operational Handler provides a **CLI** for managing embeddings, updating assistants configuration, and invoking AI-powered functions in your cluster. ### **Available CLI Operations** Install required libs (use of virtual environments recommended): ```bash pip install -r requirements.txt ``` Then list of available operations by running: ```bash python main.py --help ``` For more detail see: 1. **Invoke AI Functions** – [Functions invocation](https://github.com/Alquimia-ai/documentation/blob/main/runtime/apps/README.md) 2. **Manage Embeddings** – [Embeddings operations](https://github.com/Alquimia-ai/documentation/blob/main/runtime/embeddings/README.md) 3. **Configure Assistants** – [Configuration](https://github.com/Alquimia-ai/documentation/blob/main/runtime/config/README.md) *** ## Contributing We are building **an open, collaborative community**. Contributions are always welcome! If you'd like to add features, improve documentation, or suggest enhancements: 1. Fork the repository. 2. Create a new branch (`git checkout -b feature-xyz`). 3. Submit a **pull request** with your proposed changes. *** ## License **Alquimia Operational Handler** is open-source and available under the **MIT License**. ***