Installation

Install AEGIS and choose your integration method — Python SDK, REST API, or self-hosted source install.

AEGIS evaluates engineering proposals against quantitative governance gates. Choose your integration method:

Choose Your Integration

Method	Install	Import	Best For
Python SDK	`pip install aegis-governance`	`from aegis import Aegis`	Python apps calling the hosted API
REST API	No install needed	curl / httpx / fetch	Any language, no Python dependency
MCP	`claude mcp add aegis`	AI agent tool calling	Claude Code, Cursor, VS Code

Python SDK (Recommended)

Install the thin client SDK from PyPI:

pip install aegis-governance

from aegis import Aegis

client = Aegis(api_key="uk_live_xxx")  # or set AEGIS_API_KEY env var
decision = client.evaluate(
    proposal_summary="Add Redis caching layer",
    estimated_impact="medium",
    risk_baseline=0.1,
    risk_proposed=0.3,
)
print(decision.status)  # "proceed" | "pause" | "halt" | "escalate"

See the Python SDK Quickstart for the full guide.

REST API (No Installation)

Call the hosted API directly from any language or platform:

curl -X POST https://aegis-api-980022636831.us-central1.run.app/evaluate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "proposal_summary": "Add Redis caching layer",
    "risk_proposed": 0.3,
    "risk_baseline": 0.1,
    "complexity_score": 0.6,
    "quality_score": 0.8,
    "novelty_score": 0.4
  }'

See the REST API Quickstart for the full guide.

Self-Hosted / Source Install (Enterprise)

The full AEGIS engine is licensed under BSL 1.1. Source access is available for Enterprise customers.

git clone <your-aegis-repo>
cd aegis-governance

# Development install (includes test tooling)
pip install -e ".[dev]"

# Verify installation
pytest tests/ -v --tb=short

Requirements

Python 3.9 or later
No runtime dependencies for core functionality

Optional Dependency Groups

Install additional capabilities as needed:

# With scipy for utility z-score calculations
pip install -e ".[engine]"

# With Prometheus metrics exporter
pip install -e ".[telemetry]"

# With MCP server for AI agent integration
pip install -e ".[mcp]"

# With AWS KMS envelope encryption
pip install -e ".[kms]"

# With HSM (PKCS#11) envelope encryption
pip install -e ".[hsm]"

# Full install (all optional dependencies)
pip install -e ".[all]"

Group	Purpose	Key Packages
`engine`	Utility z-score computation	scipy
`telemetry`	Prometheus metrics exporter	prometheus_client
`config`	YAML configuration loading	pyyaml
`mcp`	MCP server support	pyyaml, mcp
`crypto`	BIP-340 Schnorr signatures	btclib, coincurve
`pqc`	Post-quantum (ML-DSA, ML-KEM)	liboqs-python
`persistence`	Workflow state storage	sqlalchemy, asyncpg
`kms`	AWS KMS envelope encryption	boto3
`hsm`	HSM PKCS#11 envelope encryption	python-pkcs11

CLI

The CLI is included with the source install:

pip install -e ".[dev]"
aegis version

Verify Installation

from aegis_governance import __version__, pcw_decide
print(f"AEGIS {__version__} installed successfully")

aegis version

On this page