Home

AgentEase Wiki

Welcome to the AgentEase wiki.

AgentEase is a secure Python SDK for structured LLM workflows that:

• scrubs a defined set of sensitive values locally before model calls
• enforces strict typed outputs with Pydantic schema validation
• supports deterministic offline runs for development and testing

Start here

• Repository: https://github.com/ejames-dev/agentease
• PyPI package: https://pypi.org/project/agentease/
• README: https://github.com/ejames-dev/agentease#readme

Core workflows (v0.2.0)

• client.triage — support-ticket triage
• client.lead_qualification — B2B lead qualification
• client.document_classification — internal document classification

Quick install

pip install agentease==0.2.0
# or
uv add agentease==0.2.0

Minimal usage

from agentease import AgentEase

client = AgentEase.offline()
result = client.triage.run("Customer was charged twice.")
print(result.model_dump_json(indent=2))

Notes

AgentEase outputs are advisory and should not replace human review, access controls, or compliance policy.

---

Wiki pages

1. Installation and environment setup
2. Offline vs live providers
3. Guardrails and PII masking behavior
4. Template contracts and response schemas
5. Release and contribution workflow