Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.jabrod.com/llms.txt

Use this file to discover all available pages before exploring further.

​
RAG Pipelines

A Pipeline in Jabrod represents a complete end-to-end configuration for processing and retrieving data. Instead of manually wiring up chunkers, embedding models, and vector databases in your own code, you configure a Pipeline in the Jabrod dashboard.

​
Pipeline Components

Every pipeline consists of:
  1. Chunking Strategy: How documents should be split into smaller pieces (Fixed size, Sentence, Semantic, etc.)
  2. Embedding Model: The AI model used to convert text chunks into vector embeddings (e.g., text-embedding-3-small).
  3. Vector Store: Where the embeddings are stored (e.g., Upstash).
  4. Retrieval Config: Rules for how data is fetched (e.g., Top K results, minimum similarity thresholds).

​
How it works

  1. You create a pipeline in the dashboard.
  2. You add Data Sources (files, URLs, text) to the pipeline.
  3. Jabrod automatically applies your configured chunking strategy to the data.
  4. Jabrod calls the configured embedding model for every chunk.
  5. Jabrod stores the embeddings in an isolated namespace within the vector database (Upstash).
  6. You generate an API Key for the pipeline.
  7. You use the API key to call the /rag/query endpoint from your application.

​
Deployment Targets

When creating a pipeline, you can specify its intended Deployment Targets. This helps organize and authorize how your RAG pipeline is consumed across your ecosystem. Jabrod supports the following targets:
  • REST API (Free / Pro): Direct API endpoint for programmatic access.
  • Website Chatbot (Pro): Embeddable chatbot widget for your website.
  • Slack & Discord (Pro): Deploy as a bot for team collaboration or community support.
  • Internal Tools (Pro): Integrate into company internal dashboards.
  • Mobile App & Voice Assistant (Pro): SDK for iOS, Android, and voice platforms like Alexa.
  • Coding Copilot & Workflow (Pro): Power AI coding assistants or use as nodes in automation workflows.
  • Enterprise Automation (Pro): Full enterprise automation and custom integrations.