Status note (2026-05-18): This repository is preserved as a public vitrine of the aXIOM project at the 2026-05-18 scope-cut. No further commits are planned here. "Ongoing methodology development takes place outside this repository, in the mentalpinG-Ops organisation on GitHub. For inquiries: mentalpingoperator@gmail.com"
For the current active/deferred track split see
STATUS.md. The content below the Historical README content line predates the scope-cut and is preserved as historical record.
This is a public snapshot of the aXIOM project at the time of its 2026-05-18 scope-cut. aXIOM is a methodological project organised around the Declarative Frame Schema (DFS) — an apparatus for AI-assisted analysis of cultural and textual artefacts.
The repository archives the product/pilot branch of the project:
- an early HTML proof-of-concept (
legacy/— Artefact Analyser v1.0), - an institutional onboarding wizard for an Academic Assessor module
(
axiom/), - supporting architecture, methodology, and discovery documents
(
docs/), - corresponding governance files (
CONTRIBUTING.md,PROMPT_GOVERNANCE.md,SECURITY_AUDIT.md, etc.).
This branch is deferred. The repository receives no further commits.
The methodology track — DFS development, conceptual work, academic contact lines — continues outside this repository. If the deferred product/pilot branch ever reactivates, work would resume in a separate repository.
| File | What it tells you |
|---|---|
STATUS.md |
Current active/deferred track split, reactivation conditions, vitrine intent. |
HISTORY.md |
Conceptual evolution timeline from Q1-2026 to the scope-cut. |
CHANGELOG.md |
Structural and scope-level changes. |
docs/METHODOLOGY.md |
Analytical methodology snapshot. |
docs/ANALYTICAL_COMMITMENTS.md |
Epistemic output commitments. |
docs/README.md |
Per-file track index for docs/. |
SECURITY_AUDIT.md |
Privacy/security audit (findings deferred). |
ARCHITECTURE.md |
Architecture document (pre-scope-cut). |
The content below predates the 2026-05-18 scope-cut and is preserved verbatim as historical record. It describes the project as it was framed before the methodology track and the product/pilot track were formally decoupled. Treat it as a snapshot of the original framing, not as a current description.
aXIOM is an AI-assisted analysis platform for academic research and teaching. It is built around four modules that share a common analytical engine.
The teacher reads, challenges, and decides. The tool carries the analytical overhead.
Module 1 — The Engine The shared analytical core. Handles content extraction, structured analysis, claim taxonomy, and inference logic. Used by all other modules.
Module 2 — Philosophical Text Analysis Analyses philosophical texts for argument structure, presuppositions, conceptual distinctions, and internal coherence. Designed for researchers in philosophy and applied social science.
Module 3 — Cultural Artefact Analysis Analyses cultural artefacts — film, media, political speech, visual material — using semiotic, ideological, and discourse analysis frameworks. Supports Zotero-based research workflows.
Module 4 — Academic Assessor Helps academic teachers assess whether student work meets the specific requirements of their department. Requirements are configured once and reused across multiple submissions. Produces a structured assessment report for the teacher.
This project is currently in a validation phase. The primary goal is not reach, but analytical usefulness and methodological integrity.
- Evidence-linked findings: every analytical claim must be traceable to specific passages or segments in the source material.
- Separation of layers: the engine distinguishes between extraction/description (what is in the text) and interpretation (a reading of what it means).
- Uncertainty is explicit: outputs include confidence/uncertainty markers and make room for alternative readings where reasonable.
- It is not an objective grader or a truth machine.
- It does not replace human interpretation. The human researcher/teacher remains responsible for critique and synthesis.
If you have domain experience in cultural studies, discourse analysis, semiotics, or qualitative research methods, feedback is most useful on:
- Whether the proposed categories and output structure support scholarly review and disagreement (counter-readings), rather than hiding it.
- Whether the tool's separation between description and interpretation is clear enough to prevent methodological misuse.
Phase: Architecture + Proof of Concept
- Architecture defined. No application code written yet.
- Modules 2 and 3 are in active personal use via the original
single-file HTML tool (see
legacy/). - Module 4 is the primary build target for phase 1.
- Pilot programme in preparation with academic institutions in Poland.
The legacy/ folder contains the original proof of concept — the
Artefact Analyser — a single-file HTML tool implementing the
analytical engine via AI prompting. It is functional and in active use,
but does not constitute a validated scientific methodology. See
METHODOLOGY.md for discussion of this boundary.
| File / Folder | Description |
|---|---|
ARCHITECTURE.md |
System architecture and decision log |
CLAUDE.md |
Anthropic Claude Code project context |
COMMS.md |
Communication strategy for stakeholder-facing materials |
CONTRIBUTING.md |
Contribution guidelines and coding standards |
FEEDBACK.md |
Feedback intake log — routes items to project or comms tracks |
.github/copilot-instructions.md |
GitHub Copilot project context |
legacy/ |
Original proof of concept (Artefact Analyser v1.0) |
docs/ |
Supporting documents (see below) |
| File | Description |
|---|---|
BIBLIOGRAPHY.md |
Reference literature |
DISCOVERY_PROTOCOL.md |
Protocol for pilot discovery sessions |
DISCOVERY_SESSION_001.md |
Notes from first discovery session |
GITHUB_ISSUES_SETUP.md |
Issue label and workflow conventions |
METHODOLOGY.md |
Analytical methodology and its current limitations |
ANALYTICAL_COMMITMENTS.md |
Epistemic/analytical output commitments |
- Python + FastAPI
- PostgreSQL
- Plain HTML / CSS / JavaScript
- Docker
No application code has been written yet. These are architectural decisions documented in ARCHITECTURE.md.
aXIOM — 2026