The CRE Acquisition Orchestrator is presented as the first comprehensive open-source AI-native framework specifically for commercial real estate acquisitions. It aims to transform traditional acquisition processes—such as due diligence, underwriting, financing, legal, and closing—into structured, executable agent workflows. The framework features 31 defined AI roles, including orchestrators, acquisition specialists, and document ingestion roles, all managed through markdown prompts.
Key functionalities include document-first deal intake, supporting various file types like rent rolls, T12s, offering memos, and PDFs. It offers source-backed extraction, where data from supported files is converted into candidate fields with confidence scores, warnings, and provenance details. A human approval gate ensures that underwriting inputs are not changed until an operator approves or waives ambiguous values. The system also provides a visible coordination dashboard to track specialist messages, handoffs, dependencies, reviews, and workpaper status.
The project emphasizes a local-first approach, allowing users to run the dashboard without API keys and process local documents. An optional path for live agents using models like ChatGPT/Codex is available, but the offline demo serves as the primary proof path. The latest release, v2.7.0, includes enhancements such as improved document intelligence with text-based PDF extraction and page-level provenance, better Excel parsing for complex workbooks, and review-grade workpapers with quality gates and evidence completeness checks. It also features live agent hardening with retry/backoff mechanisms and a single-operator self-host deployment option.