Alpha Terminal MCP Server

Alpha Terminal is the domain-specific MCP server for board-level AI governance. Built by Alpha, The AI Governance Company.

What It Does

Enables any AI system to:

• Understand board composition - Directors, committees, AI expertise, independence
• Evaluate AI governance maturity - The Alpha Standard dimension scoring under The Alpha Standard
• Detect governance risks - Red flags, governance debt, regulatory exposure
• Benchmark companies - Peer comparison, sector rankings, percentile analysis
• Generate board-ready outputs - Briefs, memos, gap analyses, defense assessments

Data Coverage

Data	Count
Companies	9,700+
Directors	2,200+
SEC Filings	345
Governance Ratings	56+ (growing)
AI Regulations	50+ jurisdictions

Quick Start

Install From PyPI

pip install alpha-terminal-mcp

For clients that should run the published package without a global install:

uvx --from alpha-terminal-mcp alpha-terminal-mcp

Configure

Set a tenant API key issued by Alpha:

export ALPHA_API_KEY="ak_live_..."

The server defaults to the production Alpha API Gateway. Override it only for local development or staging:

export ALPHA_API_BASE_URL="https://api.alpha.ac/functions/v1/api-gateway/v1"

The package also reads a local .env file:

cd alpha_mcp_server
cp .env.example .env
# Edit .env with your Alpha tenant API key

Run

# stdio mode (for Claude Desktop, Claude Code, etc.)
alpha-terminal-mcp

# Backward-compatible alias for existing clients
alpha-governance-mcp

# Or run directly
python -m alpha_governance_mcp.server

Local Development

cd alpha_mcp_server
pip install -e .

Install From Smithery

npx -y smithery@latest mcp add alpha-ac/alpha-terminal-mcp \
  --client claude \
  --config '{"apiKey":"ak_live_..."}'

Use another supported client name in place of claude when installing to Claude Code, Codex, Cursor, VS Code, or another Smithery-supported client.

MCP Client Configuration

Use this config after installing the package with pip install alpha-terminal-mcp:

{
  "mcpServers": {
    "alpha-terminal": {
      "command": "alpha-terminal-mcp",
      "env": {
        "ALPHA_API_KEY": "ak_live_..."
      }
    }
  }
}

Or use the PyPI package through uvx without a global install:

{
  "mcpServers": {
    "alpha-terminal": {
      "command": "uvx",
      "args": ["--from", "alpha-terminal-mcp", "alpha-terminal-mcp"],
      "env": {
        "ALPHA_API_KEY": "ak_live_..."
      }
    }
  }
}

To point a client at staging or local infrastructure, add ALPHA_API_BASE_URL inside env:

{
  "mcpServers": {
    "alpha-terminal": {
      "command": "alpha-terminal-mcp",
      "env": {
        "ALPHA_API_KEY": "ak_live_...",
        "ALPHA_API_BASE_URL": "https://api.alpha.ac/functions/v1/api-gateway/v1"
      }
    }
  }
}

Publish to Smithery

The remote MCP endpoint is the deployed Supabase Edge Function:

npx -y smithery@latest mcp publish \
  "https://api.alpha.ac/functions/v1/mcp-server" \
  -n alpha-ac/alpha-terminal-mcp \
  --config-schema ./smithery.config.schema.json

The Smithery config schema forwards the user-provided Alpha key as x-api-key.

For the local Python stdio package, build and publish an MCPB bundle:

uv lock
npx -y @anthropic-ai/mcpb@latest validate ./manifest.json
npx -y @anthropic-ai/mcpb@latest pack . ./dist/alpha-terminal-mcp-0.1.1.mcpb
npx -y smithery@latest mcp publish ./dist/alpha-terminal-mcp-0.1.1.mcpb -n alpha-ac/alpha-terminal-mcp

Publish to PyPI

PyPI publishing uses GitHub Actions trusted publishing. Configure the PyPI project alpha-terminal-mcp with:

• Owner: wolfepereira
• Repository: alpha-platform
• Workflow: publish-alpha-terminal-mcp.yml
• Environment: pypi

Then publish a GitHub release or run the workflow manually from main.

Claude Desktop Configuration

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "alpha-terminal": {
      "command": "alpha-terminal-mcp",
      "env": {
        "ALPHA_API_KEY": "ak_live_..."
      }
    }
  }
}

Claude Code Configuration

Add to .claude/settings.json:

{
  "mcpServers": {
    "alpha-terminal": {
      "command": "alpha-terminal-mcp",
      "env": {
        "ALPHA_API_KEY": "ak_live_..."
      }
    }
  }
}

MCP Resources

URI	Description
alpha://company/{ticker}/board	Board composition
alpha://company/{ticker}/committees	Committee structure
alpha://company/{ticker}/ai-governance-profile	Full governance profile
alpha://company/{ticker}/risk-score	AI governance risk score
alpha://company/{ticker}/proxy/latest	Latest parsed DEF 14A
alpha://regulations/{region}	Regulatory intelligence
alpha://benchmark/{sector}	Sector benchmarking

MCP Tools

Tool	Description
get_board_profile	Full board composition and director profiles
parse_def14a	Parse latest DEF 14A proxy filing
score_ai_governance_risk	Score governance risk (Alpha Governance Scale)
compare_companies	Compare governance across 2-10 companies
find_ai_governance_signals	Detect AI governance signals in filings
get_regulatory_intelligence	Regulatory obligations and timelines
generate_board_brief	Board-ready governance brief
benchmark_board_ai_readiness	Peer benchmarking and percentile ranking
search_companies	Search by name or ticker
get_platform_stats	Platform data coverage statistics

MCP Prompts (Workflows)

Prompt	Description
ai_governance_board_brief	Comprehensive board brief
proxy_change_analysis	Proxy filing change detection
audit_committee_ai_risk_memo	Audit committee risk memo
director_gap_analysis	Board expertise gap analysis
activist_defense_readiness	Activist defense assessment

Architecture

MCP Client (Claude, etc.)
    |
    v
Alpha Terminal MCP Server (FastMCP)
    |
    v
Alpha API Gateway (tenant API key)
    |
    v
PostgreSQL (alpha.ac production database)
    - companies (9,700+)
    - directors (2,200+)
    - board_seats (committee assignments)
    - governance_ratings (Alpha Governance Scale + D1-D12)
    - edgar_filings (SEC EDGAR)
    - ai_regulations (regulatory intelligence)

Methodology

All ratings are issued under The Alpha Standard by the Alpha Institute for AI Governance.

• Alpha Governance Scale: AAA to D (plus NR)
• The Alpha Standard: 8 governance dimensions scored independently
• Alpha Materiality Profile: 4-dimension materiality assessment
• Alpha Integrity Screen: Negative screening with ceiling logic
• Public scores: Capped at BB+ (59) - public data cannot verify implementation
• Assessed scores: Full AAA-D range via alpha.ac/assessment

License

Proprietary. Copyright Alpha, The AI Governance Company.