CoCounsel Standalone MCP
MCP server enabling natural language queries across CoCounsel customer insights data sources, including Gong calls, NPS feedback, skills feedback, support cases, Teams messages, and SharePoint VOC documents.
README
CoCounsel Standalone — Customer Insights Tool MCP
FastAPI backend + MCP server for CoCounsel Standalone (Core/Web) customer insights data sources.
Architecture
┌─────────────────────────────────────────────────┐
│ Claude Desktop │
│ - User conversation interface │
│ - Synthesizes final answers from theme data │
└────────────────┬────────────────────────────────┘
│ MCP Protocol (stdio)
▼
┌─────────────────────────────────────────────────┐
│ MCP Server (mcp_server_standalone.py) │
│ - 11 MCP tools (6 source-specific + 5 utility) │
│ - HTTP client to FastAPI backend │
│ - Port 8766 │
└────────────────┬────────────────────────────────┘
│ HTTP (localhost:8766)
▼
┌─────────────────────────────────────────────────┐
│ FastAPI Backend (cit/api/main.py) │
│ - 6 data pipeline modules │
│ - Snowflake Cortex Search integration │
│ - OpenArena WebSocket LLM gateway │
│ - SSE streaming for real-time progress │
└─────────────────────────────────────────────────┘
Data Sources (6)
| # | Source | Service | Filter | Citation URLs |
|---|---|---|---|---|
| 1 | Gong Calls | COCOUNSEL_VOC_SEARCH_SVC |
None (shared, contextual) | ✅ CALL_RECORDING |
| 2 | NPS | COCOUNSEL_NPS_SEARCH_SVC |
NPS_SOURCE="General" |
❌ Dashboard only |
| 3 | Skills Feedback | COCOUNSEL_SKILLS_SEARCH_SVC |
None (Standalone-only) | ❌ Dashboard only |
| 4 | Cases | COCOUNSEL_CASES_SEARCH_SVC |
ESC_PRODUCT_TITLE_C != "CoCounsel Drafting" |
❌ Case ID only |
| 5 | Teams Messages | COCOUNSEL_TEAMS_SEARCH_SVC |
CHANNEL_NAME IN ("Core Product Feedback", "Core Product Support") |
✅ WEB_URL |
| 6 | SharePoint VOC | COCOUNSEL_SHAREPOINT_SEARCH_SVC |
None (Standalone-only) | ✅ FILE_URL |
Quick Start
1. Install Dependencies
cd cocounsel_standalone_mcp
python3 -m venv .venv
source .venv/bin/activate # or .venv\Scripts\activate on Windows
pip install -r requirements.txt
pip install -r requirements-mcp.txt
2. Configure Environment
cp .env.example .env
# Edit .env with your Snowflake & ESSO credentials
3. Start Backend
./run_standalone.sh
Verify:
curl http://127.0.0.1:8766/health
# Expected: {"status":"ok","product":"standalone"}
4. Configure Claude Desktop
Edit ~/.config/Claude/claude_desktop_config.json (Linux/Mac) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"cit-standalone": {
"command": "/absolute/path/to/.venv/bin/python",
"args": ["/absolute/path/to/cocounsel_standalone_mcp/mcp_server_standalone.py"],
"env": {
"CIT_API_BASE": "http://127.0.0.1:8766",
"CIT_MCP_TIMEOUT": "200",
"SNOWFLAKE_ACCOUNT": "a206448_prod.us-east-1",
"SNOWFLAKE_USER": "your.email@thomsonreuters.com",
"SNOWFLAKE_PAT": "your-token",
"ESSO_TOKEN": "your-esso-token",
"WS_WORKFLOW_ID": "04d8d46e-0806-4653-b8ba-e718d919c567"
}
}
}
}
Restart Claude Desktop (quit completely, not just close).
5. Test
Click the tools icon (hammer) in Claude Desktop — you should see 6+ CIT tools:
search_gong_datasearch_nps_datasearch_skills_feedback_datasearch_cases_datasearch_teams_datasearch_sharepoint_datacit_health_check
Test prompt:
Check the health of the CIT system.
MCP Tools
Source-Specific Tools (6)
| Tool | Description |
|---|---|
search_gong_data |
Sales & customer call transcripts |
search_nps_data |
NPS feedback (General/Standalone) |
search_skills_feedback_data |
Skills feedback (Pendo) |
search_cases_data |
Support cases (excludes Drafting) |
search_teams_data |
Teams Core channel feedback |
search_sharepoint_data |
Customer interviews & VOC docs |
Utility Tools
| Tool | Description |
|---|---|
cit_health_check |
Check backend health |
Test Prompts
Single Source Queries
1. "What are the top pricing objections in Gong calls about CoCounsel Core?"
2. "What are the main NPS complaints for CoCounsel Standalone?"
3. "Which skills are users rating lowest?"
4. "What are the most common support cases for CoCounsel Core?"
5. "What feedback came through Core Product Feedback channel this month?"
6. "What themes emerged from VOC customer interviews?"
Multi-Source Query
7. "Compare NPS feedback and Gong calls about ease of use for CoCounsel Standalone."
Project Structure
cocounsel_standalone_mcp/
├── cit/
│ ├── api/
│ │ ├── main.py # FastAPI app entry point
│ │ ├── schemas.py # Pydantic request/response models
│ │ └── routers/
│ │ └── query.py # Query router with SSE streaming
│ ├── pipelines/
│ │ ├── standalone_gong.py # Gong calls (no filter)
│ │ ├── nps.py # NPS (filtered: General)
│ │ ├── skills.py # Skills (Standalone-only)
│ │ ├── cases.py # Cases (filtered: exclude Drafting)
│ │ ├── standalone_teams.py # Teams (filtered: Core channels)
│ │ └── standalone_sharepoint.py # SharePoint (Standalone-only)
│ ├── services/
│ │ ├── snowflake_conn.py # Snowflake connection management
│ │ ├── legacy.py # Cortex Search wrapper
│ │ ├── llm_gateway.py # OpenArena WebSocket LLM client
│ │ └── utils.py # Shared pipeline utilities
│ ├── core/
│ │ ├── errors.py # Custom exceptions
│ │ └── events.py # SSE event helpers
│ ├── config.py # Standalone-specific constants
│ └── settings.py # Environment variable loader
├── mcp_server_standalone.py # MCP server (stdio transport)
├── requirements.txt # Python dependencies
├── requirements-mcp.txt # MCP-specific dependencies
├── .env.example # Environment template
├── run_standalone.sh # Start script
└── claude_desktop_config_standalone.json # Claude Desktop config template
Troubleshooting
Tools don't appear in Claude Desktop
- Check Claude Desktop was fully quit and relaunched
- Verify config JSON is valid (use jsonlint.com)
- Check paths are absolute and correct
- Run server manually to see errors:
cd /path/to/cocounsel_standalone_mcp source .venv/bin/activate CIT_API_BASE=http://127.0.0.1:8766 python mcp_server_standalone.py
Cannot connect to CIT API
- FastAPI server (step 3) must be running
- Confirm:
curl http://127.0.0.1:8766/health - Check
CIT_API_BASEin Claude Desktop config matches port
Query timeout
- Increase
CIT_MCP_TIMEOUTin Claude Desktop config (e.g.,"240") - Default is 200 seconds
Support
Team: Product Analytics
Yugendar Sairam — RelangiyugendarSaiRam@thomsonreuters.com
Ashish Khurana — Ashish.Khurana@thomsonreuters.com
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.