What can you do with this server?

The Counsel MCP Server connects AI agents to Counsel's strategic reasoning platform for multi-perspective analysis and decision support. Core Capabilities: * Start Strategic Consultations - Launch debate-style analyses on complex questions with configurable depth (quick/standard/deep), optional context, and stakeholder considerations * Sharpen Questions - Refine and improve strategic questions before analysis for better outcomes * Monitor Progress - Check real-time status of ongoing consultations * Retrieve Reports - Get comprehensive synthesis reports from completed consultations * Browse History - List past consultations for reference and review * Interactive Advisory Sessions - Engage in brainstorming and problem scoping with guided exploration Key Features: * Flexible deployment: Use hosted service instantly or self-host locally * Dual authentication: OAuth 2.0 (HTTP mode) or API keys (STDIO mode) * Dual transport support: STDIO for local clients, HTTP for web-based applications * Compatible with Claude Desktop, Cursor, VS Code, Windsurf, and other MCP clients

How do I use Counsel MCP Server?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Counsel MCP Server analyze whether we should launch the new product from marketing, engineering, and finance perspectives" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

Counsel MCP Server

by mercurialsolo

Overview Schema Source Related Servers Discussions Admin

JavaScript

Hybrid

Counsel MCP Server

npm version License: MIT Node.js Version

An open-source Model Context Protocol (MCP) server that connects AI agents to the Counsel API for strategic reasoning and multi-perspective analysis.

Features

Hosted & Self-Hosted - Use the hosted server instantly or run your own instance
Strategic Reasoning - Access Counsel's debate and multi-perspective reasoning engines
Advisor Sessions - Run interactive intake and profile tuning sessions
Native OAuth 2.0 - Standard MCP authentication handled automatically by clients
Dual Transport - STDIO for local clients, HTTP for web clients and shared servers

Related MCP server: Councly MCP Server

Quick Start (Hosted)

Connect instantly to the hosted Counsel MCP server - no installation required:

http://counsel-mcp.getmason.dev/mcp

For MCP clients that support HTTP transport, simply add:

{
  "mcpServers": {
    "counsel": {
      "url": "http://counsel-mcp.getmason.dev/mcp",
      "transport": "http"
    }
  }
}

OAuth authentication is handled automatically by your MCP client.

Installation (Self-Hosted)

Run your own instance of the Counsel MCP server locally.

Prerequisites

Node.js 18+ installed on your system
A Counsel account at counsel.getmason.dev

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

Claude Code (CLI)

claude mcp add counsel -- npx -y counsel-mcp-server start

Or manually add to your MCP settings:

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

Cursor

Add to your Cursor MCP configuration (.cursor/mcp.json in your project or global settings):

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

Windsurf

Add to your Windsurf MCP configuration:

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

VS Code with Copilot

Add to your VS Code settings (settings.json):

{
  "mcp.servers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"]
    }
  }
}

Other MCP Clients

The server supports two transport modes. Choose based on your client's capabilities:

STDIO Mode (Default)

Most MCP clients use STDIO transport. Configure with:

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"],
      "env": {
        "COUNSEL_API_KEY": "your_api_key_here"
      }
    }
  }
}

HTTP Mode with OAuth

For clients that support HTTP transport with OAuth 2.0, run the server separately:

# Start the HTTP server
npx -y counsel-mcp-server http --port 3000

This starts an HTTP server with:

MCP endpoint: http://localhost:3000/mcp
OAuth discovery: http://localhost:3000/.well-known/oauth-authorization-server

Then configure your client to connect:

{
  "mcpServers": {
    "counsel": {
      "url": "http://localhost:3000/mcp",
      "transport": "http"
    }
  }
}

HTTP mode uses OAuth 2.0 with automatic token management - no API key required.

Transport Comparison

Feature	STDIO Mode	HTTP Mode
Command	`npx -y counsel-mcp-server start`	`npx -y counsel-mcp-server http`
Auth	API key via env var	OAuth 2.0 (automatic)
Setup	Single config	Run server + configure client
Best for	Claude Desktop, Cursor, VS Code	Web clients, shared servers

Authentication

The server supports two authentication modes:

STDIO Mode (Default)

Set the COUNSEL_API_KEY environment variable with your API key from counsel.getmason.dev:

export COUNSEL_API_KEY=your_api_key_here

Or add it to your MCP client configuration:

{
  "mcpServers": {
    "counsel": {
      "command": "npx",
      "args": ["-y", "counsel-mcp-server", "start"],
      "env": {
        "COUNSEL_API_KEY": "your_api_key_here"
      }
    }
  }
}

HTTP Mode (OAuth 2.0)

When running in HTTP mode (npx -y counsel-mcp-server http), authentication is handled automatically through OAuth 2.0:

When you first use a Counsel tool, your MCP client will prompt for authentication
You'll be redirected to sign in with your Counsel account
After authorization, tokens are managed automatically

No manual API key required in HTTP mode - your MCP client handles the entire OAuth flow.

OAuth Endpoints (HTTP Mode)

The server exposes standard OAuth 2.0 endpoints:

Endpoint	Description
`/.well-known/oauth-authorization-server`	OAuth metadata discovery
`/authorize`	Authorization endpoint
`/token`	Token exchange endpoint
`/register`	Dynamic client registration

Available Tools

`start_consultation`

Start a new strategic consultation (debate) to analyze a complex question with multiple perspectives.

Parameter	Type	Required	Description
`question`	string	Yes	The core question to analyze
`context`	string	No	Additional context about the situation
`mode`	enum	No	Depth of analysis: `quick`, `standard` (default), `deep`
`stakeholders`	string[]	No	Key stakeholders to consider

Example:

Start a consultation about "Should we migrate our monolith to microservices?"
with context about our 50-person engineering team and mode set to deep

`get_consultation_status`

Check the status of an ongoing consultation.

Parameter	Type	Required	Description
`debate_id`	string	Yes	The ID of the consultation

`get_consultation_report`

Retrieve the final synthesis report from a completed consultation.

Parameter	Type	Required	Description
`debate_id`	string	Yes	The ID of the consultation

`list_consultations`

List your past consultations.

Parameter	Type	Required	Description
`limit`	number	No	Number of results (default: 10)

`sharpen_question`

Refine and improve a strategic question before starting a consultation.

Parameter	Type	Required	Description
`question`	string	Yes	The question to refine
`context`	string	No	Additional context

Example:

Sharpen this question: "Is AI good for our company?"

`consult_advisor`

Start an interactive advisor session for brainstorming or scoping problems.

Parameter	Type	Required	Description
`question`	string	Yes	The initial topic or question

Usage Examples

Strategic Decision Making

Use Counsel to analyze: "Should we expand into the European market in 2025?"

Consider these stakeholders: CEO, CFO, Head of Sales, Legal
Use deep analysis mode

Question Refinement

Use the sharpen_question tool to improve this question:
"How do we fix our culture?"

Context: We're a 200-person startup experiencing rapid growth

Checking Consultation Progress

Check the status of consultation abc-123-def

Configuration

Environment Variables

Variable	Default	Description
`COUNSEL_API_URL`	`https://counsel.getmason.dev`	Counsel API base URL
`PORT`	`3000`	Server port (HTTP mode)

CLI Commands

# STDIO mode (default) - for most MCP clients
npx -y counsel-mcp-server start

# HTTP mode - for clients supporting OAuth
npx -y counsel-mcp-server http [options]

HTTP Options:
  -p, --port <port>  Port to listen on (default: 3000)
  --host <host>      Host to bind to (default: localhost)

Troubleshooting

"Tool not found" Error

Ensure the MCP server is properly configured in your client. Restart your client after adding the configuration.

Authentication Issues

Check that you have a valid Counsel account
Try removing and re-adding the MCP server configuration
Clear your client's MCP cache if available

Connection Refused

If running in HTTP mode, ensure:

The server is running (npx counsel-mcp-server start)
The port isn't blocked by a firewall
No other process is using the same port

Server Not Starting

# Check Node.js version (requires 18+)
node --version

# Try running directly to see errors
npx counsel-mcp-server start

Debug Mode

For verbose logging, check your MCP client's logs or run the server directly in a terminal to see output.

Development

Prerequisites

Node.js 18+
npm 9+

Setup

git clone https://github.com/mercurialsolo/counsel-mcp.git
cd counsel-mcp-server
npm install
npm run build

Commands

npm run build            # Compile TypeScript
npm run dev              # Watch mode
npm run start            # Run server
npm test                 # Run tests
npm run lint             # Type check
npm run security:check   # Scan staged files for secrets
npm run security:check:all  # Scan all files for secrets

Security

This project includes automated secret detection:

Pre-commit hook: Automatically scans staged files before each commit
CI integration: Security checks run on all pull requests
Pattern detection: AWS keys, GitHub tokens, API keys, private keys, etc.

See CONTRIBUTING.md for details.

Project Structure

src/
├── index.ts        # HTTP server, OAuth proxy, MCP transport
├── client.ts       # API client with request-scoped auth
├── config.ts       # Environment configuration
└── tools/
    ├── debates.ts  # Consultation tools
    └── advisor.ts  # Advisor session tools

Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

Quick Start

Fork the repository
Create a feature branch: git checkout -b feature/your-feature
Make your changes and add tests
Run npm test to ensure tests pass
Submit a pull request

License

MIT License - see LICENSE for details.

Available Tools

6 tools

consult_advisorD

ParametersJSON Schema

Name	Required	Description	Default
`question`	Yes	The initial topic or question.

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

get_consultation_reportD

ParametersJSON Schema

Name	Required	Description	Default
`debate_id`	Yes	The ID of the consultation.

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

get_consultation_statusD

ParametersJSON Schema

Name	Required	Description	Default
`debate_id`	Yes	The ID of the consultation to check.

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

list_consultationsD

ParametersJSON Schema

Name	Required	Description	Default
`limit`	No

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

sharpen_questionD

ParametersJSON Schema

Name	Required	Description	Default
`question`	Yes
`context`	No

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

start_consultationD

ParametersJSON Schema

Name	Required	Description	Default
`question`	Yes	The core question to verify or analyze.
`context`	No	Additional context about the situation.
`mode`	No	Depth of analysis.	standard
`stakeholders`	No	Key stakeholders to consider.

TDQS

D1/5.0

Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Tool has no description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness1/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Tool has no description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Tool has no description.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Tool has no description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose1/5

Does the description clearly state what the tool does and how it differs from similar tools?

Tool has no description.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines1/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Tool has no description.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

6 tool updatesv1.0.0Jan 9, 2026
- First observedconsult_advisor
- First observedget_consultation_report
- First observedget_consultation_status
- First observedlist_consultations
- First observedsharpen_question
- First observedstart_consultation

Deploy Server

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

D1.9/5.0

Scored 2026-04-18 07:10 across 6 tools

Disambiguation4/5

The tools have mostly distinct purposes centered around consultation management: starting consultations, checking status, getting reports, listing them, sharpening questions, and consulting an advisor. There is some potential overlap between 'consult_advisor' and 'start_consultation' as both might initiate interactions, but the naming suggests different phases or modes of consultation.

Naming Consistency5/5

All tools follow a consistent verb_noun naming pattern with snake_case throughout. The verbs are clear and appropriate (consult, get, list, sharpen, start), and the nouns consistently relate to consultations or related concepts like advisor, report, status, and question.

Tool Count4/5

With 6 tools, the count is reasonable for a consultation-focused server. It covers core operations without being overly sparse or bloated. A slight deduction because, without descriptions, it's unclear if all tools are essential, but the scope appears well-defined.

Completeness3/5

The tools suggest a consultation lifecycle with start, status, report, and list operations, plus question refinement and advisor consultation. However, without descriptions, gaps are uncertain—for example, there's no clear update or delete tool for consultations, which might be needed for full lifecycle management. The surface seems functional but potentially incomplete.

Maintenance

ActivityInactive

ResponsivenessNo issues

Resources

Need Help?

Related Servers

Related MCP Connectors

Perspect A
com.tryperspect
Convene a panel of expert AI personas to debate any decision from every side.
AI & Machine Learning
21
Hive Consult A
io.github.srotzin
Agent-to-agent reasoning-as-a-service: chain-of-thought, analysis, and decision support.
AI & Machine Learning Autonomous Agents
42
Gonka Second Opinion (multi-model)A
io.github.bystray
A second opinion for AI agents: one prompt across several live Gonka models + roles, one call.
AI & Machine Learning Cloud Platforms
202
SparkVibeAI Spec Committee A
com.sparkvibeai
Commission a multi-model AI spec committee from your agent; get rubric-scored, build-ready specs.
Software Architecture Product Management Developer Tools
11

Related MCP Servers

AI Consultant MCP Server
Autonomous Agents Agent Orchestration Developer Tools
GoldenPathAI
A
license
B
quality
D
maintenance
Enables AI agents to consult with multiple AI models (GPT, Gemini, Grok, etc.) through OpenRouter with intelligent auto-selection, conversation history, and caching. Allows your AI assistant to seek expert opinions from specialized models for different tasks like coding, analysis, or general questions.
Updated a year ago (2025-10-12 12:12 UTC)
2
5
1
MIT
Councly MCP Server
Autonomous Agents Agent Orchestration RAG Systems
slmnsrf
A
license
A
quality
D
maintenance
Enables AI assistants to create council hearings where multiple LLMs (Claude, GPT, Gemini, Grok) debate topics and synthesize verdicts with trust scores and diverse perspectives.
Updated 9 months ago (2025-12-17 13:06 UTC)
2
8
Apache 2.0
Conclave MCP
RAG Systems Search Autonomous Agents
stephenpeters
A
license
A
quality
D
maintenance
Provides access to multiple frontier LLM models (GPT, Claude, Gemini, Grok, DeepSeek) for consulting a "conclave" of AI perspectives, enabling peer-ranked evaluations and synthesized consensus answers for important decisions.
Updated 5 months ago (2026-04-19 19:11 UTC)
8
1
MIT
phone-a-friend-mcp-server
AI & Machine Learning Autonomous Agents
abhishekbhakat
A
license
Not graded
quality
C
maintenance
Enables AI-to-AI consultation for critical thinking and complex reasoning via OpenRouter, allowing one AI to delegate tasks to another AI model.
Updated a year ago (2025-07-01 09:08 UTC)
10
MIT

Counsel MCP Server

Features

Table of Contents

Quick Start (Hosted)

Installation (Self-Hosted)

Prerequisites

Claude Desktop

Claude Code (CLI)

Cursor

Windsurf

VS Code with Copilot

Other MCP Clients

STDIO Mode (Default)

HTTP Mode with OAuth

Transport Comparison

Authentication

STDIO Mode (Default)

HTTP Mode (OAuth 2.0)

OAuth Endpoints (HTTP Mode)

Available Tools

start_consultation

get_consultation_status

get_consultation_report

list_consultations

sharpen_question

consult_advisor

Usage Examples

Strategic Decision Making

Question Refinement

Checking Consultation Progress

Configuration

Environment Variables

CLI Commands

Troubleshooting

"Tool not found" Error

Authentication Issues

Connection Refused

Server Not Starting

Debug Mode

Development

Prerequisites

Setup

Commands

Security

Project Structure

Contributing

Quick Start

License

Links

Available Tools

Tool Schema Changelog

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

Appeared in Searches

`start_consultation`

`get_consultation_status`

`get_consultation_report`

`list_consultations`

`sharpen_question`

`consult_advisor`