Transform Unstructured Data for AI Applications

An open-source ingestion pipeline designed to transform unstructured data into vectorized formats for integration with knowledge bases

Splinter UI
Automated Deployment to <strong>AWS</strong> with Splinter <strong>CLI</strong> Tool

Automated Deployment to AWS with Splinter CLI Tool

Splinter is built using AWS architecture and can be easily deployed using the CLI tool. The self-managed, serverless design makes it especially useful for organizations seeking both scalability and control over their unstructured data pipelines.

Data Pipeline <strong>Observability</strong>

Data Pipeline Observability

The dashboard offers real-time updates during the ingestion process, including the current state of each document’s ingestion, the number of existing vectors, and the number of new vectors added to the knowledge base.

<strong>RAG</strong> Sandbox

RAG Sandbox

The built-in RAG system uses the user’s destination database as its knowledge base. It updates its context based on the documents in the knowledge base, allowing users to verify the accuracy of the ingested data through direct queries.

<strong>Event-Driven</strong> and <strong>Ephemeral</strong> Architecture

Event-Driven and Ephemeral Architecture

Splinter synchronizes data sources like AWS S3 or Dropbox with the destination database, automatically updating whenever documents are added, updated, or removed. Splinter’s ephemeral, event-driven design ensures that it "scales to zero," eliminating costs for idle compute resources.

Meet the Team

Amir Sadeghifar

Amir Sadeghifar

Software Engineer

Miami, FL

Brian Ouyang

Brian Ouyang

Software Engineer

Los Angeles, CA

Harold Camacho

Harold Camacho

Software Engineer

Toronto, Canada

Ricardo Delgado

Ricardo Delgado

Software Engineer

Houston, TX