Open Swarm / Feature film

RAG engine — Put your knowledge to work.

1:26 · English captions

In this film

The story

Your assistant becomes more useful when it can work with your information. Open Swarm retrieves relevant material from the knowledge you have made available. The knowledge engine stores passages, finds relevant context for a question, and supplies it to an authorized assistant. This is retrieval, without retraining the model.

About the demonstrationCross-app films use illustrative chains, not recordings of a completed end-to-end run. Access checks still apply. Ambient listening is off by default; external speech services may process audio. Print to RAG uses recovered text and review, not an unverified image-OCR promise.
Read the full transcript

0:00Your assistant becomes more useful when it can work with your information. Open Swarm retrieves relevant material from the knowledge you have made available.

0:10The knowledge engine stores passages, finds relevant context for a question, and supplies it to an authorized assistant. This is retrieval, without retraining the model.

0:21Bring in reference text through supported applications. Print to RAG adds an inbox where you review documents and approve their knowledge destinations.

0:31Keyword retrieval works with stored text. A configured embedding model adds semantic matching. Results retain document and source information so the context can be inspected.

0:44Use private, shared, or application-specific knowledge as supported by your installation. Current caller permissions and declared application access control what can be retrieved.

0:56Jarvis and connected application concierges can use the knowledge they are authorized to read, keeping useful context close to the work.

1:05Run the knowledge store, models, and processing locally to keep the workflow on your own installation. Connected providers receive the content you send.

1:15Your documents. Your context. More useful assistance. Open Swarm.