Axiora MCP Server

A Model Context Protocol server that gives AI agents access to financial data for 4,000+ Japanese listed companies. Search companies, analyze financials, read translated filings, screen stocks, track ownership changes, and more — all from Claude, Cursor, or any MCP-compatible client.

Data is sourced from EDINET (Japan's FSA) — audited XBRL filings, not scraped or AI-estimated.

Free API key: axiora.dev — full access to all companies and all endpoints.

Available Tools (33)

Company Discovery

Tool	Description
search_companies	Find companies by name (JP/EN), securities code, or EDINET code
search_companies_batch	Look up multiple companies at once (up to 10)
get_company	Detailed info for a single company
get_sector_overview	List all 33 TSE sectors or get stats for one sector

Financial Data

Tool	Description
get_financials	Revenue, net income, assets, equity, ROE, ROA, margins — up to 20 years
get_growth	YoY growth rates and 3yr/5yr CAGRs
get_health_score	Financial health score (0-100) with component breakdown and risk flags
get_peers	Find peer companies in same sector with similar revenue
compare_companies	Side-by-side financials for 2-5 companies
get_timeseries	Chart-ready time-series for any metric across multiple companies

Screening & Rankings

Tool	Description
screen_companies	Filter by sector, min revenue, min ROE, max PE, and more
get_ranking	Rank all companies by any metric (revenue, ROE, margins, etc.)
get_health_ranking	Rank by financial health score

Filing Content & Translations

Tool	Description
get_sections	Full Japanese text from annual filings (MD&A, risk factors, governance...) with English translations
get_translations	English translations of a specific filing
search_translations	Full-text search across all translated filings

Filings & Coverage

Tool	Description
list_filings	List filings with filters (company, doc type)
get_filing_calendar	Filing calendar for a month (filings per day)
get_coverage	Data coverage stats (totals, freshness)

Buybacks

Tool	Description
get_buybacks	Monthly buyback execution reports — shares acquired, amounts spent, completion %

Ownership Intelligence

Tool	Description
get_shareholdings	Large shareholding reports (>=5% holders) with purposes and funding sources
get_ownership_trajectories	Per-filer ownership time-series showing accumulation/exit patterns
get_ownership_movers	Market-wide biggest ownership moves ranked by velocity
get_ownership_signals	Detected signals: accumulation streaks, step-ups, exits, activist escalations
get_cross_holdings	Cross-holding pairs where two companies hold >=5% of each other
get_probability_table	P(reaching 10%/20%/33%
get_capital_allocation	Classify a company as Returner, Hoarder, Reinvestor, or Mixed
get_capital_allocation_ranking	Rank companies by capital allocation style
get_activist_campaigns	Detected activist campaigns with holding trajectories and outcomes
get_unwinding_scoreboard	Cross-holdings being dissolved — tracking Japan's governance reform

Governance

Tool	Description
get_board_composition	Directors, auditors, officers with outside/independent status and gender breakdown
get_voting_results	AGM voting results — proposal-level votes, approval %, outcomes

Setup

Prerequisites

• Python 3.10+
• uv package manager

Installation

1. Clone this repository:

   git clone https://github.com/axiora-dev/mcp-server.git
   cd mcp-server

2. If you don't have uv installed:

   curl -LsSf https://astral.sh/uv/install.sh | sh

3. Set up your API key:

   cp .env.example .env
   # Edit .env and add your API key from https://axiora.dev

4. Run the server:

   uv run server.py

Client Configuration

Claude Desktop

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

{
  "mcpServers": {
    "axiora": {
      "command": "uv",
      "args": ["--directory", "/absolute/path/to/mcp-server", "run", "server.py"],
      "env": {
        "AXIORA_API_KEY": "your-api-key"
      }
    }
  }
}

Claude Code

claude mcp add axiora -- uv --directory /absolute/path/to/mcp-server run server.py

Cursor

Add to Cursor Settings → MCP Servers:

{
  "mcpServers": {
    "axiora": {
      "command": "uv",
      "args": ["--directory", "/absolute/path/to/mcp-server", "run", "server.py"],
      "env": {
        "AXIORA_API_KEY": "your-api-key"
      }
    }
  }
}

VS Code (Copilot)

Add to .vscode/mcp.json in your project:

{
  "servers": {
    "axiora": {
      "command": "uv",
      "args": ["--directory", "/absolute/path/to/mcp-server", "run", "server.py"],
      "env": {
        "AXIORA_API_KEY": "your-api-key"
      }
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "axiora": {
      "command": "uv",
      "args": ["--directory", "/absolute/path/to/mcp-server", "run", "server.py"],
      "env": {
        "AXIORA_API_KEY": "your-api-key"
      }
    }
  }
}

Example Prompts

Once connected, try asking your AI assistant:

• "What are Toyota's financials for the last 5 years?"
• "Compare Sony, Nintendo, and Bandai Namco"
• "Find Japanese companies with ROE above 15% in the electronics sector"
• "What risks does SoftBank mention in their latest annual report?"
• "Show me the healthiest companies in the information & communication sector"
• "Who is accumulating shares in Japanese companies right now?"
• "Show me activist campaigns in Japan"
• "What's Toyota's board composition?"
• "Which companies are classified as capital returners?"

Alternative: Hosted MCP (Zero Install)

Don't want to run a local server? Axiora also offers a hosted MCP endpoint at https://api.axiora.dev/mcp with the same tools — no installation required.

Python SDK

For programmatic access beyond MCP, install the Python SDK:

pip install axiora

from axiora import Axiora

client = Axiora(api_key="your-api-key")
toyota = client.companies.retrieve("7203")
financials = client.companies.retrieve_financials("7203", years=10)
results = client.screen.retrieve(min_roe=15.0, sector="電気機器")

Data Coverage

Dimension	Detail
Companies	~4,000 listed + ~7,500 non-listed (11,000+ total)
Metrics	52 per company per fiscal year
History	Up to 15 years (FY2010-FY2025)
Standards	JP GAAP, IFRS, US GAAP (auto-normalized)
Translations	English translations of MD&A, risk factors, governance, and more
Ownership	19,000+ large shareholding reports with trajectory analysis
Buybacks	6,900+ monthly buyback execution reports
Updates	Every 30 minutes from EDINET
Source	Audited XBRL filings (金融庁 Financial Services Agency)

Links

• Axiora — Dashboard, API keys, docs
• API Documentation
• Python SDK on PyPI

License

MIT