Data Room
Upload · parse · cite — fuel for the AI thesis layer
Upload research
Annual reports · transcripts · investor decks · brokerage PDFs · policy documents.
Document library
5 documents · all chunks tagged by company, segment and topic
vector store: pgvector (mock)
| Document | Type | Pages | Uploaded | Status |
|---|---|---|---|---|
| Waaree Energies — FY25 AR.pdf | Annual Report | 312 | 2026-04-22 | Parsed |
| Premier Energies — Q4 FY25 Earnings Call.txt | Earnings Call | 28 | 2026-05-04 | Parsed |
| MNRE — ALMM List-II Cell Notification.pdf | Government PDF | 11 | 2026-03-09 | Parsed |
| SECI Tranche XV — RFP.pdf | Tender Document | 96 | 2026-04-17 | Parsing |
| Adani Green — Investor Day Deck.pptx | Investor Presentation | 72 | 2026-02-12 | Parsed |
AI thesis generator
Generates one-line · bull · bear · risk · monitorables · final view, cited to documents.
Step 1
Extract
OCR + chunk + embed every uploaded doc into pgvector. Metadata: company, year, segment, topic.
Step 2
Retrieve
For a company query, retrieve top-k chunks across ARs, transcripts and policy PDFs, ranked by topic relevance.
Step 3
Generate
Run the 16-section institutional thesis prompt with mandatory citations to chunk IDs.