How anmolsanger/real_estate_ai-agent works
EstateMind accepts a supported Indian market and property type, runs a multi-agent research and analysis workflow, then combines the analyst’s judgment with Python-verified financial metrics in a report. Users can initiate runs through the Streamlit UI, FastAPI, or CLI. The README documents these interfaces and the research/modeler/analyst stages; sampled source confirms the pipeline’s finance, persistence, and telemetry responsibilities. Some API and CLI details are represented only by the README and source index.
Main entry points: pipeline.py Analysis pipeline · app.py Streamlit UI · api.py FastAPI · storage.py Artifact storage
Components 21
User interfaces
- Streamlit UI
app.py - FastAPI
api.py - CLI
crew.py - City validation
cities.py - MCP finance server
mcp_server.py
Analysis workflow
- Request and report models
schemas.py - Analysis pipeline
pipeline.py
Research agents
- Property researcher
agents.py - Financial modeler
agents.py - Investment analyst
agents.py - Agent task prompts
tasks.py - Web search
tools.py
Financial modeling
- ROI calculator tool
tools.py - Deterministic finance
finance.py
Reports and operations
- Verified report
pipeline.py - Artifact storage
storage.py - Run telemetry
telemetry.py
Other components
- Investor
- Property listing sites
- S3-compatible storage
- Local artifacts
Connections 27
- Investor → Streamlit UI: submits analysis
README.md - Investor → FastAPI: requests analysis
README.md - Investor → CLI: starts analysis
README.md - Streamlit UI → City validation: selects city
app.py - Streamlit UI → Request and report models: builds assumptions
app.py - Streamlit UI → Analysis pipeline: runs analysis
app.py - FastAPI → Request and report models: validates request
api.py - FastAPI → Analysis pipeline: runs analysis
api.py - FastAPI → Run telemetry: reports run data
api.py - CLI → Analysis pipeline: runs analysis
crew.py - Analysis pipeline → Property researcher: creates agent
pipeline.py - Analysis pipeline → Financial modeler: creates agent
pipeline.py - Analysis pipeline → Investment analyst: creates agent
pipeline.py - Analysis pipeline → Agent task prompts: creates tasks
pipeline.py - Analysis pipeline → Deterministic finance: validates and scores
pipeline.py - Analysis pipeline → Request and report models: builds result
pipeline.py - Analysis pipeline → Artifact storage: saves artifacts
pipeline.py - Analysis pipeline → Run telemetry: records run
pipeline.py - Property researcher → Web search: uses search
agents.py - Web search → Property listing sites: searches listings
- Financial modeler → ROI calculator tool: calculates candidates
agents.py - ROI calculator tool → Deterministic finance: averages and calculates
tools.py - Investment analyst → Verified report: writes judgment
README.md - Analysis pipeline → Verified report: adds verified metrics
pipeline.py - Artifact storage → S3-compatible storage: uploads artifacts
storage.py - Artifact storage → Local artifacts: falls back to
storage.py - MCP finance server → Deterministic finance: exposes formulas
mcp_server.py