Rayven Generative AI Node (LLaMA): Secure, Private Generative AI for Any Application

Rayven.io provides a fully integrated Generative AI Node powered by a LLaMA-based language model running entirely on Rayven-managed or customer-hosted infrastructure.

This allows you to use large language models (LLMs) securely within Rayven workflows and applications—without relying on third-party providers like OpenAI.

This node enables dynamic AI-driven outputs across use cases, while ensuring data privacy, control, and compliance with data sovereignty requirements.


Key Capabilities

1. On-Platform LLM Execution

Rayven’s Generative AI runs locally on Rayven’s servers or can be deployed in your private cloud or on-premise. No API calls to external services are required.

  • No third-party data exposure

  • Compliant with privacy and sovereignty standards

  • Suitable for internal, regulated, or sensitive applications

2. Prompt-Based AI Logic

You can create and manage prompts within the Generative AI Node. Prompts can include inputs from:

  • Real-time workflow values

  • Table records

  • Uploaded files (PDFs, CSVs, plain text)

  • External data ingested through connectors

3. Workflow Integration

The AI Node is a native component in Rayven’s low-code workflow builder. Use it to:

  • Process inputs dynamically

  • Generate AI-based responses

  • Route outputs to logic, dashboards, alerts, or storage

4. Private Cloud Deployment Option

Rayven supports dedicated deployment of the LLaMA AI engine to your private infrastructure:

  • Full ownership of the model

  • Complete control over hosting, access, and usage

  • Suitable for enterprises with strict IT governance


Common Use Cases

Automated Summaries in Workflows

Use the AI Node to summarize daily operations, shift activities, or status logs:

Prompt: “Summarize this shift’s performance: [insert workflow variables]”
Output: “Production met 94% of target. Minor delays due to equipment resets. No critical faults reported.”


Document-Based Analysis

Upload an internal policy, SOP, or log file and generate a concise summary or response:

Prompt: “Summarize the main procedures described in this PDF.”
Output: “The document outlines emergency response steps, focusing on containment, isolation, and reporting.”


Proposal or Email Generation

Use workflow or table data to automatically create proposals, quotes, or email content:

Prompt: “Create a maintenance proposal based on the following usage metrics: [insert data]”
Output: “We recommend replacing Filter A every 2 weeks and scheduling monthly inspections for Pump 3.”


App Interface Personalization

Insert AI-generated text into dashboards or forms:

  • HTML nodes display human-like recommendations

  • Table columns show predicted next actions

  • Widgets update with live, LLM-generated insights


Workflow Output Destinations

Generative AI outputs can be sent to:

  • Tables: Write structured results alongside existing data

  • HTML Nodes: Display free-text insights or summaries

  • Email/SMS Nodes: Send personalized, AI-generated content

  • APIs or Webhooks: Forward results to external systems


Benefits

  • No External Dependencies: Everything runs inside your secured Rayven or private infrastructure

  • Prompt Flexibility: Tailor prompts using any input data source

  • Real-Time Responses: AI logic runs inline with workflow execution

  • Scalable Integration: Use in any workflow, app, or output node

  • Compliance-Ready: Designed for organizations with data governance or sovereignty requirements


Q&A

Q: Is Rayven’s Generative AI connected to OpenAI or external models?
A: No. Rayven uses a self-hosted LLaMA-based model that runs entirely on Rayven or customer infrastructure. No data leaves your environment.


Q: Can I use workflow and table data inside prompts?
A: Yes. You can dynamically insert real-time workflow values, historical records, or user input directly into the prompt to generate customized outputs.


Q: Can the AI Node read files?
A: Yes. You can upload documents (PDF, CSV, text) into the workflow, then reference their contents in prompts for summarization or analysis.


Q: Where can the AI output be used?
A: You can send outputs to dashboards, HTML widgets, tables, alerts (email/SMS), or use them to inform decisions within the workflow itself.


Q: Can I host the LLM privately?
A: Yes. Rayven offers a private cloud or on-premise deployment of the LLaMA model, giving you full control over data, hosting, and access.


Q: Does this require coding?
A: No. The Generative AI Node is used within Rayven’s no-code workflow builder. Prompts and outputs are configured visually.


Q: Is the output updated in real time?
A: Yes. Each time the workflow runs, the prompt is executed using the latest available data, ensuring outputs are always current.


Q: Can I use multiple AI nodes in one workflow?
A: Yes. You can use multiple Generative AI nodes for different prompts or use cases—e.g., one for summaries, another for recommendations.