🔍 AISA Compliance Checker

Evaluate your agentic AI project against the AISA reference architecture.

Upload your project as a ZIP file and get an instant compliance report across 7 architectural layers and 4 cross-layer contracts.

Layer	What's Checked
L1: LLM Foundation	Model adapters, prompts, context management, safety
L2: Tool & Environment	Tool schemas, sandboxing, permissions, MCP
L3: Cognitive Agent	Planning, memory, goals, reflection
L4: Infrastructure	Orchestration, state, multi-agent, tracing
L5: Evaluation	Tests, monitoring, regression, human eval
L6: Dev & Deployment	Versioning, CI/CD, A/B testing, rollout
L7: Governance	Policy-as-code, privacy, fairness, oversight

CLI version: pip install aisa-checker

Upload & Configure

Upload Project (ZIP)

Analysis Mode

Pattern scan is free. AI modes require an API key.

Pattern Scan (Fast) AI Analysis Full Analysis (Pattern + AI)

How to prepare your ZIP:

ZIP your project's root directory
Include all Python files, configs, CI/CD files
Exclude node_modules, .venv, .git (they're auto-skipped)

Results

Upload a project and click 'Run AISA Check' to see results.

AISA (Agentic AI Systems Architecture) — A unified layered reference architecture for designing, deploying, evaluating, and governing agentic AI systems.

Nacar, O., Alquffari, D., & Alkhalifa, M. (2026). Tuwaiq Academy — TRDC.

Paper & Resources | PyPI Package | Mapping Study