HN Showcase
Back to browse
I parsed 10 years of Japanese corporate filings into an API

I parsed 10 years of Japanese corporate filings into an API

by dahaleonkar·Mar 2, 2026·4 points·0 comments
Visit ProjectView on HN

AI Analysis

●●●●GemZero to OneSolve My ProblemBig Brain

Unlocked Japanese financial data—XBRL parsed to JSON for 11k companies, no signup.

Strengths
  • •Genuine zero-to-one: Japanese financial APIs didn't exist; XBRL parser solves a real, deep friction (3 accounting standards, company taxonomies, Japanese-only legacy system).
  • •Production-grade: typed Python SDK (Stainless), OpenAPI spec, MCP integration, keyset pagination, webhooks—not a scraper, not a hack.
  • •Coverage is comprehensive and current: 11k companies, 10 years, 52 fields, 30-min updates; free to start without login.
Weaknesses
  • •Geographic lock-in: valuable only for Japan-focused investors; limits addressable market compared to U.S. SEC or global data providers.
  • •Roadmap dependent on TDNet real-time feeds; current version is historical filings only.
Category
Finance
Target Audience

Quants, fintech engineers, researchers analyzing Japanese equities; no signup required to start.

Similar To

Alpha Vantage · Finnhub · SEC Edgar API

Post Description

Hi HN,

Japan is the world's 4th largest stock market, but its financial data is stuck in the 90s. Corporate filings go through EDINET — a government system that stores everything as XBRL in Japanese, with 3 different accounting standards (JP-GAAP,IFRS, US-GAAP), company-specific taxonomies, and no usable API.

I've been working with Japanese XBRL filings for 2 years. I built a parser that normalizes all of it into structured JSON. Axiora now covers 4,125 listed companies with ~10 years of history and 52 financial fields per company-year.

Built for developers:

· REST API with full OpenAPI spec — https://axiora.dev/docs · Python SDK: pip install axiora · LangChain integration: pip install langchain-axiora (20 tools) · MCP server for Claude, ChatGPT, Cursor, etc. · llms.txt at axiora.dev/llms.txt

Try it now, no signup: https://axiora.dev/terminal

Roadmap: TDNet integration (real-time timely disclosures from JPX) and stock prices are next. Data licenses are expensive and I'm bootstrapping this solo, so those will come as paid tiers.

Tech: Python/FastAPI on Cloud Run, Supabase (Postgres), lxml for XBRL parsing.

Free tier: 10,000 req/day. I'd love feedback on the API design or what data would be most useful.

Similar Projects

Productivity●●Solid

I treated my CV like a data product-evidence.json,MCP endpoint,llms.txt

Evidence-mapped CV beats PDF for AI recruiter parsing, but applies only to ATS that read these formats.

Big BrainShip It
vassilbek
103mo ago