1 points | by takeshi_sakamo 9 hours ago
1 comments
PolyAgora is an experiment in building a structured multi-agent reasoning system entirely through natural language — no code, no scripts.
It evolved into a 6-agent cognitive architecture (Arc / Ann / Saku / Kanzaki / Yui / Kou), each representing an orthogonal reasoning vector.
The goal isn’t to create “characters,” but to test whether an OS-level reasoning frame can emerge from constraints written purely in natural language.
Happy to explain the architecture or the constraint design.
PolyAgora is an experiment in building a structured multi-agent reasoning system entirely through natural language — no code, no scripts.
It evolved into a 6-agent cognitive architecture (Arc / Ann / Saku / Kanzaki / Yui / Kou), each representing an orthogonal reasoning vector.
The goal isn’t to create “characters,” but to test whether an OS-level reasoning frame can emerge from constraints written purely in natural language.
Happy to explain the architecture or the constraint design.