Workflow All workflows API reference

Equity research — one stock, four lenses

Composed flow across concall_search · co_ratios · sector_get · /companies/{X}/holders. The pattern: read what management says, check valuation trajectory, frame against peers, confirm institutional positioning. Worked through with HINDALCO as the running example.

Tools used · concall_search (search) · co_ratios (read) · sector_get (read) · /companies/{X}/holders (search)
Total credit cost · ~12 credits per stock for the full flow

export GENKA_API_KEY=imk_live_...
H="X-API-Key: $GENKA_API_KEY"
S=HINDALCO    # the stock we're researching

1. What is management actually talking about?

Lead with concalls -- the narrative that the numbers will or will not catch up to. Symbol-scoped search returns this company's full transcript trail.

curl -sH "$H" "https://genka.dev/latest/concalls/search?q=margin&symbol=$S&limit=5"

For HINDALCO this surfaces commentary on aluminium spread compression, Novelis (US subsidiary) margin trajectory, and the renewables-pivot capex ramp. Pair with q=capex&symbol=$S for the spend cycle and q=novelis&symbol=$S for the US story specifically. Each hit returns a snippet so you read the management voice, not a summary.

2. Is the valuation expanding or contracting?

Quarterly P/E time-series — not the latest snapshot — tells you whether the market is paying up or punishing.

curl -sH "$H" "https://genka.dev/latest/companies/$S/ratios?metric=pe"

Returns a points[] array with {period_end, value} for every disclosed quarter since FY17. HINDALCO Q4FY25 9.58 → Q1FY26 9.20 → Q2FY26 9.64 → Q3FY26 12.39 — +35% off the trough in two quarters. That's the multiple expanding while you weren't watching. Cross-reference: was Q3 the quarter management started talking up Novelis? Step 1's concall search will tell you.

3. Who are the peers, and how does this stock compare?

SEBI sectoral classification + aggregate ratios. Frames the stock against its actual cohort instead of the index.

curl -sH "$H" "https://genka.dev/latest/sectors/non-ferrous-metals?expand%5B%5D=members&expand%5B%5D=aggregate"

Returns members[] (the cohort) and aggregate (median P/E, ROE, market cap across the sector). Compare HINDALCO's ratios against the sector median: if the company P/E is below sector median but earnings growth is above, that's an asymmetric setup. co_screener with the same sector + a roe_gte filter narrows further.

4. Who already owns it, and how convicted are they?

Reverse-index from a stock to the funds holding it — the institutional positioning lens. Concentration tells you whether smart money has already moved.

curl -sH "$H" "https://genka.dev/latest/companies/$S/holders?limit=10" | jq '.data | sort_by(-.weight_pct)'

Returns the funds owning HINDALCO ranked by weight. 251 funds for JINDALSTEL as a comparison — if the top 5 are 4-8% positions, that's high conviction. If everyone's at 1%, it's index-tracking, not active conviction. /holders/summary aggregates total INR exposure across all funds — useful for "how much MF capital is at risk if this stock breaks?".

5. What did the analyst questions miss?

Q&A in concalls is the cleanest signal of what the buy-side is worried about. Run the same concall_search with terms tied to specific risks.

curl -sH "$H" "https://genka.dev/latest/concalls/search?q=Novelis+margin&symbol=$S"
curl -sH "$H" "https://genka.dev/latest/concalls/search?q=aluminium+spread&symbol=$S"

For each, read the snippet field — if the same question is asked four quarters in a row and the answer mutates, that's a building issue. Pivot to concall_answer when you want a synthesised paragraph instead of fragment hits: "What did HINDALCO management say about Novelis margins in FY26?" with citations linked back via fragment id.

Same pattern works for any NSE 500 symbol. Swap $S at the top — the four other calls compose unchanged. Other workflows: portfolio review (what you already own) · macro watching (FII/DII + concall macro) · discovery (find ideas you don't know about yet).