DART Corpus DB
1.3 GB
Structured business-report chunks used for company evidence lookup, theme discovery, and FTS keyword search.
- Rows 243,897
- Companies 3,141
- Years 2021-2025
KR DART Research MCP
ONLINE
What This Is
This is a Model Context Protocol server for Korean listed-company research backed by DART filings.
The browser page is only a status view. ChatGPT connects through the SSE endpoint below.
Endpoints
MCP SSE stream: /sse
Status JSON: /status.json
Data Artifacts
243,897DART business chunks
3,141listed companies covered
2021-2025filing year range
98.3%vector map coverage
FAISS Vector Index
1.4 GB
Semantic-search index over DART chunks. It lets natural-language queries find related filing passages.
- Backend FAISS
- Use semantic search
- Tool search_dart_vector
Vector Map
1.8 MB
Maps FAISS vector rows back to DART corpus chunk ids so search results can show company and filing context.
- Mapped vectors 239,801
- Unique ids 239,801
- Coverage 98.3%
| Name | Purpose | Approx Size |
|---|---|---|
| DART corpus DB | Business-report chunks and FTS search | 1.3 GB |
| FAISS index | Semantic vector search over DART chunks | 1.4 GB |
| FAISS map | Vector row to corpus chunk id mapping | 1.8 MB |
Tools
resolve_kr_companyfind_companies_by_business_themeget_company_evidenceget_kr_financial_dataget_kr_financial_data_batchget_kr_business_evidencetrace_kr_business_changetrace_kr_business_change_by_themesearchfetchsearch_dart_vectorroute_stockanalyze_stock