Mycelium FAQs

Question 1

What output formats does Mycelium provide?

Accepted Answer

Mycelium offers various output formats tailored for different use cases: 'compact' for one-line summaries, 'digest' for structured JSON suitable for AI agents, 'manifest' for an indexed scan-and-drill-down view, and a default raw JSON 'SurvivorReport'.

Question 2

How does Mycelium classify knowledge into categories?

Accepted Answer

It runs content through a 60-tick biological ecosystem simulation where nodes compete, cooperate, and merge. The surviving content is then classified into distinct categories: pure, merged, loner, redundant, and dead, reflecting its quality and relationships.

Question 3

What is Mycelium and what does it do?

Accepted Answer

Mycelium is a bio-inspired semantic filtering engine that applies ecosystem simulations to analyze embedding vectors and text. It classifies knowledge quality based on social dynamics, identifying unique, merged, isolated, or redundant content.

Question 4

Can Mycelium identify relationships between different documents or data sources?

Accepted Answer

Yes, Mycelium features 'cross-file affinity' discovery. It can analyze semantic relationships between multiple source documents, providing an affinity matrix and stats on how content from different files merges or relates.

Question 5

Is an external Qdrant instance required for Mycelium's filtering process?

Accepted Answer

No, Mycelium performs its core filtering process entirely in-memory. While it can source data from existing Qdrant collections, it does not require a dedicated external Qdrant instance to run its simulations.

Mycelium

Mycelium

주요 기능

사용 사례

주요 기능

사용 사례