A modular, extensible platform for enterprise architecture management — EA review workflows, application portfolio management, business capability analysis, architecture decision & design (ADD), PACT concern catalog, viewpoint/artifact mapping, and reporting.
AxisRobo-PAMP was sparked by the PAMF series of papers (PACT, AVDM, AADM, ARCM) and shaped by years of enterprise architecture practice. It is built not only around what ships today, but toward a long-term vision for architecture governance — spanning domain capability coverage and platform evolution through the v3.x horizon. See the Roadmap for that vision, beyond the currently implemented feature set.
| Layer | Technology |
|---|---|
| Frontend | Next.js 16.2 (Turbopack), React 19, TanStack Query, Ant Design v6, Tailwind CSS |
| Backend | FastAPI (Python 3.12+), SQLAlchemy asyncpg, Raw SQL via text() |
| Database | PostgreSQL 14+, 100 tables in eam schema DDL |
| Auth | Pluggable: OSS local (JWT + bcrypt), Keycloak SSO (EE), Dev mode. RBAC (3 roles, 24 resources) |
| Storage | Pluggable: S3-compatible or database-backed (eam_file_storage) |
| Testing | pytest (backend + API integration), Playwright (E2E) |
- Node.js >= 22.0.0
- Python >= 3.12
- PostgreSQL 14+
# Clone
git clone <repo-url>
cd AXISRobo-PAMP
# Install frontend
cd frontend && npm install
# Setup backend
cd ../backend
python -m venv venv
source venv/bin/activate # or venv\Scripts\activate on Windows
pip install -r requirements.txt
# Configure backend
cp .env.example .envdocker run -d --name axisarch-pg \
-e POSTGRES_USER=postgres \
-e POSTGRES_PASSWORD=postgres \
-e POSTGRES_DB=axisarch \
-p 5432:5432 postgres:16The database schema and AVDM seed data are maintained as SQL scripts under docs/SQL/. Create the eam schema, load the table DDL, then load the AVDM seed data:
# Connection string matches the docker container above (db: axisarch)
PG="postgresql://postgres:postgres@localhost:5432/axisarch"
# 1. Create the eam schema
psql "$PG" -c "CREATE SCHEMA IF NOT EXISTS eam;"
# 2. Load all table definitions
psql "$PG" -f docs/SQL/eam_schema_ddl.sql
# 3. Load AVDM seed data (concerns, viewpoints, artifacts, mappings)
psql "$PG" -f docs/SQL/avdm_schema_seed.sql# Backend (port 4000)
cd backend && python -m uvicorn app.main:app --port 4000 --reload
# Frontend (port 3000)
cd frontend && npm run dev| Mode | AUTH_MODE |
Description |
|---|---|---|
| Dev | dev |
Fixed admin user (AUTH_DEV_USER), no auth required |
| Local (OSS) | local |
Username/password + JWT, users stored in eam.local_users |
| OIDC (EE) | oidc |
Keycloak SSO with JWT validation |
# backend/.env
AUTH_MODE=local
JWT_SECRET_KEY=change-me-in-production
# frontend/.env
NEXT_PUBLIC_AUTH_MODE=localDefault admin: admin / admin123
Admin can manage users at /api/users (CRUD endpoints) or via the Admin workspace.
Set AUTH_MODE=dev in backend/.env to skip authentication during development. All requests use the default admin user.
| Role | Permissions |
|---|---|
| Admin | Full access (*:*), user CRUD, all data management pages |
| Reviewer | Review assigned EA requests, manage meetings, actions, BCM |
| Requestor | Create and manage own EA requests and projects |
| Module | Description |
|---|---|
| Project Management | Project CRUD, team management |
| EA Review | Architecture review workflow, meetings, actions, AI architecture checks, AVDM questionnaire |
| ADD — Architecture Decision & Design | PACT concern catalog, viewpoint catalog, concern-to-viewpoint mapping, viewpoint-to-artifact mapping, questionnaire config, risk scoring, architecture governance matrix |
| AI Self-Assessment | AI project self-assessment with AT0-AT8 adoption tier × L0-L4 governance maturity matrix, 11-section 49-item checklist, counterparty type (CP1-CP4) governance contexts |
| Application Portfolio | BCM, BizCapability, CMDB — business capability mapping |
| Technology Management | Tech stack lifecycle, compliance checking |
| Data Management | Enterprise data governance — master data, data classification, data flow, data-application mapping, resources, certifications |
| System Management | RBAC, auth, users, audit logging |
Modules selectively enabled via ENABLED_MODULES env var.
Browser → Next.js :3000 → FastAPI :4000 → PostgreSQL :5432
→ S3 / eam_file_storage (attachments)
→ LLM API (AI architecture review)
- Modular by contract — Each module self-contained with own DB access
- Deny by default — All 95+ API endpoints RBAC-gated via
require_permission(resource, scope) - Audit everything — Append-only
eam_audit_logwithaudit_allow()/audit_deny()hooks - Plugin-first — Auth providers, email services, CMDB connectors, and storage backends are abstracted behind interfaces
- Raw SQL over ORM — SQLAlchemy
text()for performance; Pydantic for request/response validation - Storage abstraction — S3-compatible when configured, database-backed (
eam_file_storage) when not
AXISRobo-PAMP implements AVDM as a configurable decision chain:
Questionnaire Answer
→ Risk Signal / Architecture Concern Activation
→ Concern Score Aggregation
→ Viewpoint Classification
→ Artifact Recommendation
The questionnaire captures project scale, complexity, change scope, and architecture type. Selected answers activate architecture concerns through weighted mapping rules. Activated concerns are then projected onto the viewpoint taxonomy through concern-to-viewpoint mappings. Finally, each selected viewpoint is mapped to one or more supporting artifacts, such as diagrams, lists, or matrices.
Canonical chain: Questionnaire → Concern → Viewpoint → Artifact.
Multiple concerns may map to the same viewpoint, and a single concern may activate multiple viewpoints. Artifacts are derived exclusively through viewpoints — there is no direct concern-to-artifact mapping. The viewpoint layer is the single source of truth for which artifacts a concern requires.
The seed configuration includes an active cross-border-sensitive-data combination rule. It escalates data sovereignty, data compliance, security control, and compliance-boundary concerns when cross-border data flow is combined with sensitive-data or checkpoint conditions.
The walkthrough in the AVDM manuscript reports a selected activation trace that isolates convergent single-signal activation. The full rule set also supports the stronger combination-rule case when the additional checkpoint predicates are satisfied.
Activation rules come in two kinds, both stored in the same activation-rule table but distinct mechanisms:
- Single-signal — one
in/equalspredicate over a single answer. The architecture-type rules over the multiselectarchitectureTypeSection.*fields are single-signal (array-valued) rules, not combination rules. - Combination — multiple
ALL/ANYpredicates across several answers (e.g.cross-border-sensitive-data).
95+ endpoints, all RBAC-gated. Swagger UI at http://localhost:4000/docs.
See docs/api.md for full reference.
| Backend | Config | Use Case |
|---|---|---|
| S3 | Set S3_ENDPOINT |
Production, enterprise deployments |
| Database | S3 not configured (default) | Local development, OSS deployments |
Files stored in eam.eam_file_storage table when S3 is unavailable.
100 tables across the eam schema DDL. Complete documentation at docs/database-schema.md.
# Generate fresh schema documentation
cd backend && python ../scripts/generate_db_schema_doc.py# API integration tests
cd api-tests && python -m pytest
# Backend unit tests
cd backend && python -m pytest
# Frontend E2E
cd frontend && npx playwright test| Doc | Description |
|---|---|
| docs/architecture.md | System architecture |
| docs/api.md | API reference |
| docs/STATUS.md | Completion status |
| docs/ROADMAP.md | Development roadmap |
| docs/threat-model.md | Security threat model |
| docs/design.md | Detailed design (Chinese) |
| docs/database-schema.md | Complete DB schema (100 tables) |
| docs/SQL/eam_schema_ddl.sql | Complete schema DDL — CREATE TABLE definitions for all tables |
| docs/SQL/avdm_schema_seed.sql | AVDM schema + seed data (concerns, viewpoints, artifacts, mappings) |
| docs/authorization.md | Auth model: RBAC + record-level ownership |
| docs/standards/ | Coding conventions |
If you use AXISRobo-PAMP in your research, please cite the related papers:
[1] Han, H. PACT: A Reference Viewpoint Taxonomy for Software-Intensive Systems. Preprints 2026, 2026010720. DOI: 10.20944/preprints202601.0720.v1
[2] Han, H. AVDM: A Risk-Informed Decision Model for Architecture Viewpoint Selection. TechRxiv, 27 January 2026. DOI: 10.36227/techrxiv.176948759.92525674/v1
[3] Han, H. AADM: An Architecture Artifact Decision Model for Risk-Aware Documentation Scope. TechRxiv, 27 January 2026. DOI: 10.36227/techrxiv.176948904.46258539/v1
[4] Han, H. ARCM: Responsibility Configuration in Multi-View Architecture Descriptions. TechRxiv, 27 January 2026. DOI: 10.36227/techrxiv.176948938.83267225/v1
[5] Han, H. AXISRobo-PAMP: An Open-Source Enterprise Architecture Management Platform. GitHub repository, 2026. [Online]. Available: https://github.com/axisrobo/AXISRobo-PAMP
MIT
