AIDE Systems

AI systems for structured inquiry.

AIDE’s systems help human operators frame a subject, investigate a situation, develop and revise work, and carry knowledge forward.

System roles

AIDE systems serve distinct roles.

Select a system to see what it supports, when it is useful, and how it relates to the human operator.

AIDE System Map

Distinct systems within a governed inquiry.

Distinct systems support the work; human purpose, judgment, and responsibility remain visible.

Role-Specific Cognitive Support Systems

Cognitive AIDEs

Frame tasks, preserve human authority, test claims, and support useful work without replacing judgment.

View the four AIDEs Public access requirements

Maturity: Developed role-specific configurations under continuing refinement.
Example input: A reasoning, research, analytics, or writing task, beginning with an alignment-focused intake.
Example output: A task-appropriate analysis, research synthesis, data interpretation, or draft with explicit limits.
Availability: Reasoning, Research, Analytics, and Writing AIDEs are publicly available as Custom GPTs.
Evidence status: Operational use informs iteration.
Research relationship: Shared role-specific systems supporting inquiry, research, analysis, and writing across AIDE Research.
Human authority boundary: The user retains judgment, style, authorship, and final decision-making authority.

Structural Analysis Instrument

Archimedes

Represents and analyzes complex situations, examines plausible outcomes, and prepares decision surfaces that improve human judgment.

Explore Structural Analysis

Maturity: Developed analytical architecture under continuing use and refinement.
Example input: A defined problem, system, text, or phenomenon requiring relational and structural analysis.
Example output: A structured analytical synthesis describing a system's elements, relations, constraints, tensions, trajectories, supporting evidence, and structural implications.
Availability: Publicly available as a Custom GPT. Research collaboration and demonstration inquiries are also welcome.
Evidence status: Evidence Integration is part of the method; formal comparative evaluation remains future work.
Research relationship: A current implementation of Structural Analysis within Structured Reasoning; operation informs refinement and evaluation.
Human authority boundary: The instrument prepares decision-ready analysis; it does not authorize conclusions or replace accountable interpretation.

Conversational Cognitive System

Roger

AIDE’s primary Conversational Cognitive System, supporting sustained inquiry, synthesis, conceptual investigation, and architectural development while preserving operator authority.

Roger remains central to ongoing architectural and research work.

About Roger

Maturity: Developed architecture in active internal use and observation.
Example input: A complex question, relevant context, constraints, prior work, and the human user's intent.
Example output: A reasoned synthesis, clarified decision surface, or durable update to an ongoing body of work.
Availability: Private system. Demonstration or research collaboration may be considered by inquiry.
Evidence status: Active internal use and observation; public evaluation has not yet been published.
Research relationship: A shared inquiry and development environment contributing across AIDE Research; its use does not establish a research claim.
Human authority boundary: The human user retains authority over purpose, interpretation, judgment, and decisions.

Governed Knowledge Infrastructure

Knowledge Stewardship System

Maintains coherent, evolving knowledge bases through structured artifacts, continuity practices, decisions, and traceable change.

View Knowledge Stewardship

Maturity: Active internal practice with an evolving system design.
Example input: Files, claims, definitions, decisions, revisions, and research trajectories accumulated over time.
Example output: A durable, navigable knowledge base with explicit status, continuity, provenance, and governed updates.
Availability: Currently an internal AIDE capability; demonstrations and institutional applications may be discussed.
Evidence status: Active internal practice; public demonstrations are in development.
Research relationship: Shared infrastructure preserving evidence, definitions, decisions, and revisions across AIDE Research.
Human authority boundary: AI may identify and draft changes; publication and consequential changes remain subject to human approval.

Cognitive AIDEs

Every task begins with framing.

Every Cognitive AIDE begins by helping the human operator define and frame the task: the problem to be reasoned over, the bounds of research, the object of analysis, or the audience and appropriate voice for writing. Only then does it move into its role-specific work.

Public access requirements

Public Custom GPT

Reasoning AIDE

Clarifies questions, assumptions, concepts, and decision surfaces before producing an answer.

Get help understanding a problem —

Open Reasoning AIDE in ChatGPT

Public Custom GPT

Research AIDE

Investigates sources, separates evidence from interpretation, preserves uncertainty, and resists premature certainty.

Get help investigating a question or source —

Open Research AIDE in ChatGPT

Public Custom GPT

Analytics AIDE

Structures data questions, interprets results, identifies limits, and supports decisions without substituting for accountable judgment.

Get help interpreting data or a result —

Open Analytics AIDE in ChatGPT

Public Custom GPT

Writing AIDE

Drafts and revises language while preserving the author’s intent, style, authority, and final editorial control.

Get help editing and clarifying your writing —

Open Writing AIDE in ChatGPT

Knowledge stewardship

An inquiry should not have to begin again.

Knowledge Stewardship preserves the question, evidence, definitions, interpretations, decisions, and unresolved issues that let an inquiry continue across time, systems, and participants.

It gives later work a visible starting point: what has been established, what remains uncertain, and what requires revision.

Current form and evidence status

Current form and open work.

Role: Durable knowledge infrastructure for continuing inquiry.
Current form: Active internal practice, structured records, and evolving system design.
Current use: Preserving research, system development, definitions, evidence, decisions, and revision history.
Public availability: Internal capability; demonstrations and institutional applications may be discussed.
Open work: Inspectable public demonstrations and evaluation of continuity practices.
Authority boundary: AI may identify and draft changes; humans approve publication and consequential updates.

Public access and portability

Available now through ChatGPT.

Archimedes and AIDE’s four Cognitive AIDEs are currently hosted as public Custom GPTs in ChatGPT.

To use them, sign in to ChatGPT. Access and usage limits depend on your plan, region, and workspace settings.

AIDE systems have also been successfully deployed in Microsoft Copilot and with Claude as an underlying model. Those implementations are not currently public.

See Structural Analysis View Current Work