⚠️ Panoramica is archived!
As of February 2024, Panoramica has been made read-only and archived. That includes this site as well as the GitHub repository. If you want to learn about the origins and ideas behind project, you can still read the the announcement post.
Panoramica is not dead, it's just not open source anymore. It lives on as the internal backend for Orbit's Community Search product, powering search and AI features over conversational data. If you're interested in Panoramica-like features and insights for your community, please reach out and let us know. Panoramica has an API you can use.
The decision to stop pursuing the project as open source was not an easy one, but we believe that the rapid acceleration of LLM-related tools has made it possible to use general-purpose tools to achieve similar results. We are leaving the codebase up for community developers who may find inspiration in the graph data model and graph analytics use cases.
Thanks for your support!
—Josh, Steeve, and the rest of the Orbit team
🧰 What does Panoramica do?
Panoramica helps developers build AI-powered experiences on top of conversational data, from data ingestion to graph construction to interfacing with end users and AI.
1. Ingest Conversations
Panoramica pulls in messages from sources like Discord, GitHub, Discourse, and Twitter. Important metadata like actors, mentions, replies, and parent/reply relationships are included.
2. Update the Graph
Nodes are created for every message and actor. Conversation trees are built by creating edges that join parents with replies. In this format, we can do basic graph queries as well as more advanced graph algorithms like Pagerank, shortest path, and community detection.
3. Equip an AI Assistant
Panoramica finds and formats the data so that an AI assistant can correctly assist users with a wide variety of questions and tasks. Based on the user's request, the right data is retrieved from vector and graph databases.
