FCCS MCP Agentic Server
Oracle EPM Cloud Financial Consolidation and Close (FCCS) agentic server using Google ADK with MCP support.
Features
25+ FCCS Tools: Full coverage of Oracle FCCS REST API
Dual Mode: MCP server (Claude Desktop) + Web API (FastAPI)
Memory & Feedback: PostgreSQL persistence with RL tracking
Mock Mode: Development without real FCCS connection
Bilingual: English and Portuguese support
Quick Start
Windows (Recommended)
Automated Setup:
This will:
Create virtual environment
Install all dependencies
Create
.envfile from templateGuide you through configuration
Manual Setup:
Create virtual environment:
python -m venv venvActivate:
.\venv\Scripts\Activate.ps1Install:
pip install -e .Configure: Copy
.env.exampleto.envand editInitialize database:
python scripts\init_db.py(if using PostgreSQL)
Quick Commands:
Start web server:
.\start-server.batStart MCP server:
.\start-mcp-server.batInstall dependencies:
.\install-dependencies.batInitialize database:
.\init-database.bat
See WINDOWS_DEPLOYMENT.md for detailed Windows setup guide.
Linux/Mac
1. Install Dependencies:
2. Configure Environment:
3. Run:
MCP Server (for Claude Desktop):
Web Server (for API access):
Interactive CLI:
Claude Desktop Configuration
Add to %APPDATA%\Claude\claude_desktop_config.json:
API Endpoints
Endpoint | Method | Description |
| GET | Health check |
| GET | List available tools |
| POST | Execute a tool |
| POST | Call specific tool |
| POST | Submit user feedback |
| GET | Get tool metrics |
Available Tools
Application
get_application_info- FCCS application detailsget_rest_api_version- API version info
Jobs
list_jobs- List recent jobsget_job_status- Job status by IDrun_business_rule- Execute business rulesrun_data_rule- Execute data load rules
Dimensions
get_dimensions- List all dimensionsget_members- Get dimension membersget_dimension_hierarchy- Build hierarchy tree
Journals
get_journals- List journalsget_journal_details- Journal detailsperform_journal_action- Approve, reject, postupdate_journal_period- Update periodexport_journals/import_journals
Data
export_data_slice- Export grid datasmart_retrieve- Smart data retrievalcopy_data/clear_data
Reports
generate_report- Generate FCCS reportsget_report_job_status- Async report status
Consolidation
export_consolidation_rulesets/import_consolidation_rulesetsvalidate_metadatagenerate_intercompany_matching_reportimport_supplementation_datadeploy_form_template
Architecture
Deployment
Windows
See WINDOWS_DEPLOYMENT.md for complete Windows deployment guide including:
Prerequisites installation
Automated setup scripts
Windows Service configuration
Troubleshooting
Docker
Google Cloud Run
See QUICK_DEPLOY.md for detailed Cloud Run deployment.
Feedback System
The agent tracks tool executions for reinforcement learning:
Automatic: Execution time, success/failure, errors
User Feedback: 1-5 rating via
/feedbackendpointMetrics: Aggregated stats via
/metricsendpoint
Documentation
Windows Deployment Guide - Complete Windows setup
GitHub Setup Guide - Repository setup and configuration
Quick Deploy - Google Cloud Run deployment
ChatGPT Quick Start - ChatGPT integration
Dashboard Quick Start - Performance dashboard
License
MIT