proscenium
Proscenium
Proscenium is an emerging library for building collaborative enterprise AI applications.
Sticking with the implied theater analogy, the highest-level building block is a Production, which are composed of Scenes. Scenes are composed of Characters and Props.
Repository
See the repository on GitHub. Clone locally or start a new GitHub Codespace
Colab Demos
- “Abacus” to use simple tools to respond to user questions.
- “Literature” to use a vector database to store chunks of two books, and using those chunks to form a context for LLM inference based on a user question
- “Slack” to attach Proscenium Productions (Scenes of Characters and Props) to Slack channels using a custom app.
Resources
For more background and future plans, see the wiki
To find the Proscenium community, see the discussions
Use Cases
Bartlebot
For federal case law research. See the website and repo
Library Design
Proscenium is being developed with an understanding that details from specific enterprise application domains should be prioritized in the early phases of its design.
The AI model, software, hardware, and cloud ecosystem is evolving rapidly. Proscenium is designed to co-evolve with the ecosystem by keeping interfaces simple and loosely-coupled.
The highest, app-level concepts are the theater-inspired classes. Productions are composed of Scenes. Scenes are composed of Characters and Props.
Under those higher-level classes are “patterns” for systems including:
- Document Enrichment
- Knowledge Graph Construction
- Entity Resolution
- RAG
- Graph RAG
- Tool Use
At the lowest level are “verbs” such as: chunk, complete (inference), extract, and invoke (tools). In many cases these are thin wrappers around well-known libraries.