Philosophical
Research System
A multi-agent AI platform where three specialized agents collaboratively investigate any philosophical question, each operating through a distinct, user-selected intellectual framework.
How It Works
The user types any philosophical question and selects one framework per agent from a catalog of 15 intellectual methodologies. The three agents then run sequentially, each one receiving the previous agent's output as context, producing a layered, multi-perspective analysis.
Philosophical Researcher
Investigates the question through a chosen research framework, from Analytic Philosophy to Genealogy. Surfaces the strongest arguments, counter-arguments, cross-cultural perspectives, and unresolved tensions.
Epistemic Critic
Audits the researcher's report using a chosen critique framework, from Adam Smith's Impartial Spectator to Kuhn's Paradigm Analysis. Identifies framing biases, missing voices, and blind spots.
Integral Synthesizer
Integrates both reports through a chosen synthesis framework, from Wilber's AQAL to Rawls' Reflective Equilibrium. Distinguishes genuine integration from false consensus.
Framework Catalog
Technical Stack
What I Learned
Building this system taught me how to structure multi-agent pipelines where each agent has a distinct role, a specific context window, and a clear handoff to the next. Designing the framework injection system, where philosophical methodologies are encoded as detailed prompt instructions, showed me how much output quality depends on how precisely you define a role.
On the frontend side, implementing real-time Server-Sent Events (SSE) streaming and the typewriter effect required thinking carefully about state, what's happening in the browser while the server is still generating. It was the first time I experienced the full loop from a user interface decision all the way down to how a Python thread queues messages.
Most importantly, this project confirmed my core belief: AI is most powerful not when it replaces thinking, but when it structures it, giving different frameworks equal weight, holding contradictions open, and refusing easy synthesis.