Chat
Research questions about your holdings. Dossiers run on Research.
-
Session restored from this Mac. 3 dossiers on file · 8 holdings synced · core running · read-only broker sync.
-
Which of my holdings is furthest from its own fair-value base?
-
CRANE. Its filed Stage 6 range is bear ₹1,190 / base ₹1,720 / bull ₹2,240 and it last traded at ₹1,498.60 — about 12.8% below base, the widest gap among the filed ranges. ACME is the other side of that: ₹1,563.20 against a ₹1,450 base, so the price is 7.8% above its own range midpoint.
Both numbers come from filed stages, not from me. Open the dossiers on Research to see the receipts behind each range.
Live assistant is not connected in this build.
Research only. Algo Agent never places, sizes or alters a trade — the app has no order endpoint.
Portfolio
8 holdings · KITE, manual
Live 8/8 · 09:44:12 IST
TOTAL EQUITY VALUE
₹8,88,705.00+₹67,512.50 (+8.2%)
Invested
₹8,21,192.50
Holdings
8
Valued
8/8
live price available
Allocation
- ACME 21%
- BHARAT 17%
- CRANE 13%
- NEXUS 12%
- RIDGE 11%
- ORION 9%
- LOTUS 9%
- VELUM 8%
Holdings
Select a holding to open its research dossier. The chart icon opens the price sheet. Column headers sort.
| Actions | ||||||||
|---|---|---|---|---|---|---|---|---|
| 120.00 | ₹1,284.50 | ₹1,563.20 | ₹1,87,584.00 | +₹33,444.00 (+21.7%) | +1.4% | 21.1% | ||
| 75.00 | ₹1,610.00 | ₹1,498.60 | ₹1,12,395.00 | −₹8,355.00 (−6.9%) | −1.2% | 12.6% | ||
| 340.00 | ₹402.15 | ₹441.05 | ₹1,49,957.00 | +₹13,226.00 (+9.7%) | −0.6% | 16.9% | ||
| 40.00 | ₹2,220.00 | ₹2,101.75 | ₹84,070.00 | −₹4,730.00 (−5.3%) | +0.2% | 9.5% | ||
| 500.00 | ₹128.40 | ₹149.85 | ₹74,925.00 | +₹10,725.00 (+16.7%) | +2.3% | 8.4% | ||
| 150.00 | ₹612.00 | ₹688.40 | ₹1,03,260.00 | +₹11,460.00 (+12.5%) | +0.9% | 11.6% | ||
| 90.00 | ₹968.75 | ₹1,041.30 | ₹93,717.00 | +₹6,529.50 (+7.5%) | +1.1% | 10.5% | ||
| 260.00 | ₹298.40 | ₹318.45 | ₹82,797.00 | +₹5,213.00 (+6.7%) | −1.6% | 9.3% | ||
| 8 holdings | ₹8,88,705.00 | |||||||
Performance
Prices from your connected providers (Zerodha Kite · IndianAPI). Updated just now. Research-only — not advice.
MARKETS
Live · 09:44:12 ISTNSE · BSE via your Zerodha Kite feed. Day move is measured against the previous settlement; the line is the live tape since this screen opened.
Cash Balances
- INR₹48,250.00
- USD$1,204.60
Totals are reported per currency and never blended across them.
31 Aug 2026
- Kite holdings syncread-only · NSE/BSE8 equities · 2 cash buckets re-read09:41 IST
- CSV importmanual · local fileschema checked · 0 rows added, nothing overwritten08:12 IST
18 Aug 2026
- Kite session refreshtoken rotated 06:00 ISTHoldings re-read after daily expiry · no order scope requested06:04 IST
- Manual CSV importVELUM · NSE500 units @ ₹128.40 recorded from a local file14:22 IST
12 Aug 2026
- First Kite connectread-only scope4 equities pulled · trade tokens were never requested11:08 IST
Read-only sync events from your broker and from files you imported. Algoagent has no order endpoint — it cannot place, modify or cancel anything.
Research
Nine code-enforced stages. Citations are receipts of fetches the app actually made.
ACME
- 1Mandate & Identity: complete
- 2Industry & Position: complete
- 3Moat & Durability: complete
- 4Financial Quality: complete
- 5Management & Governance: complete
- 6Valuation: complete
- 7Pre-mortem & Risks: complete
- 8Catalysts & Monitorables: complete
- 9Synthesis & Verdict: complete
9 of 9 stages filed · audit passed
donerevision 3 · 9 stages filed · 212 receipts stored · audit passed
Your position
live priceAvg cost
₹1,284.50
Price
₹1,563.20
Qty
120.00
Unrealised
+21.70%
52-week high
₹1,689.90
52-week low
₹1,022.40
100-day high
₹1,612.05
200-day high
₹1,689.90
The price is 7.5% below the 52-week high, 52.9% above the low, sitting at 81% of the year's range; your entry at 39%.
Extremes from intraday highs and lows, over 248 days of bars from IndianAPI.
- Bear ₹1,180.00
- Base ₹1,450.00
- Bull ₹1,890.00
Dossier verdict
Fair
Upside to base
−7.2%
At ₹1,563.20 the price sits above base — inside the filed range but past its midpoint, with −7.2% to base.
You bought at ₹1,284.50 — below base.
Stage 6's range, not a target. Informational assessment, not investment advice.
Verdict
assessment, not adviceCONVICTION
E(R) −0.8%
Gates 5/7
Conviction 61 of 100.
QUALITY7.2/10
CONFIDENCE81%
Scored over 92 of 100 weight — 8 could not be scored and was removed from the denominator, not defaulted.
Confidence caps enforced in code: any abstained stage caps it at 60%, no primary source caps it at 40%. Neither fired here.
⚠ concentration — top-five clients are 41% of revenue and the disclosure names no replacement pipeline
Quality6.8 financials + 0.4 news14 news rows
Pros
- Durable cash generation; ROIC 18.4% against a 12.1% WACC across five years
- Clean balance sheet — debt/equity 0.18, interest coverage 14.2×
- Stable governance: no promoter pledge, same auditor four years
Cons
- Mature growth; large-client budget pressure two quarters running
- FX and discretionary-spend sensitivity in the export book
- Concentration in a few verticals
Red flags
- Receivables days up 11 to 68 over four quarters
- Any related-party transaction above the 2% threshold
- Margin compression two quarters running
Bear
Top client insources; revenue falls 9% and operating margin gives back 260bps. ₹1,180.
Base
Mid-single-digit growth, margin flat, buyback continues at the filed run-rate. ₹1,450.
Bull
Two verticals re-accelerate and the pricing test holds; multiple re-rates to the five-year mean. ₹1,890.
Core valuation
computed by the core- reverse dcf1,180.00 – 1,450.00 – 1,890.00
- earnings power value1,362.00
- peg screen1.42
- ev ebitda14.6
- dividend discountabstained — payout has not been stable for three yearsneeds: payout_ratio_3y, dps_history
Arithmetic over stored statements and the inputs the core could resolve. Informational assessment, not investment advice.
Final cross-stage audit
claude-opus-4 · independent model-
Coverage gap
Stage 5 records no CFO tenure, and Stage 7's pre-mortem leans on management continuity. The claim stands but is thinner than the pre-mortem implies.
stages.management.findings.compensation_structure · stages.premortem.findings.top_risks[1]
-
Cross-stage consistency
Stage 3 rates the moat wide on switching costs; Stage 8 lists a competitor migration tool as an upgrade trigger for that same customer base. Not a contradiction, but the two read against each other.
stages.moat.findings.moat_sources · stages.catalysts.findings.downgrade_triggers
2 unusable findings discarded — each named a field this dossier does not have, or a number it does not contain.
The auditor runs in a fresh context with its tools disabled. It can only inspect the record and object; it never re-scores.
Stages
1 · Mandate & Identity100%
Acme Industries Ltd builds and services process-automation controllers for mid-size Indian manufacturers, and takes a recurring maintenance fee on the installed base. Roughly 62% of revenue is hardware plus commissioning; 38% is the annuity contract that follows it.
Earnings model
Install once, bill the maintenance annually
Geography
India 71% · Middle East 18% · SE Asia 11%
Key products
AX-7 controller · SiteLink telemetry
2 · Industry & Position92%
FIVE FORCES
Market share
11.4%
Industry growth
7.8%
Maturity
Mature, consolidating
3 · Moat & Durability100%
Rating
Wide
Trend
Stable
Durability (years)
8
Switching cost is the source: a controller swap means re-certifying the line, and the FY25 report puts average customer tenure at 9.4 years. The 5% price-raise test was run in two verticals in FY24 with no measurable churn, which is the evidence the rating rests on.
COUNTER-EVIDENCE
- A competitor shipped a migration tool in Q2 FY26 that claims to cut re-certification to one shift. No third-party confirmation was found.
4 · Financial Quality100%
ROIC
18.4%
WACC
12.1%
ROIC vs WACC
+6.3 pts
Cash conversion
88%
Revenue CAGR 5y
9.6%
Operating margin
19.2%
Debt/equity
0.18
Interest coverage
14.2×
ACCOUNTING FLAGS
- Receivables days 57 → 68 across four quarters while revenue grew 6%. Flagged, not concluded.
5 · Management & Governance88%
Promoter pledging
0%
Insider ownership
41.2%
Auditor tenure
4 years
No auditor change in four years and no qualified opinion on file. Related-party transactions are 0.7% of revenue, below the 2% threshold the framework flags at.
Gap: CFO tenure could not be resolved from any fetched filing. Recorded as unknown rather than inferred from the signature block.
6 · Valuation100%
P/E (current)
28.4
P/E (5y avg)
24.1
PEG
1.42
EV/EBITDA
14.6
Implied growth
11.3%
Upside to base
−7.2%
The reverse DCF asks what growth the current price already assumes: 11.3% for a decade, against a filed five-year record of 9.6%. That gap is the whole valuation verdict — not expensive, but paying today for an acceleration nothing in the record has demonstrated.
WHAT WOULD CHANGE THE VERDICT
- Two consecutive quarters above 11% organic growth would make the implied number a forecast rather than a stretch.
- Receivables normalising below 60 days would raise the cash-conversion input the base case runs on.
7 · Pre-mortem & Risks100%
Assume it is 2029 and this position lost half its value. The most likely story: the largest client insourced controller maintenance, two others used that as leverage on renewal pricing, and the annuity line — the half of revenue that carries the margin — reset a full year before the hardware line could replace it.
KILL CRITERIA
- Annuity revenue declines two consecutive quarters year on year.
- Top-client concentration passes 45% of revenue.
- Any auditor change accompanied by a restatement.
Base rate: roughly one in three industrial-automation incumbents in this size band lost a top-three account within five years of a migration tool reaching the market.
8 · Catalysts & Monitorables76%
Filed with gaps — the stage names its own missing inputs rather than filling them.
MONITORABLES
- Quarterly annuity revenue disclosure (Q3 FY26, expected 22 Jan).
- Receivables days in the next two balance sheets.
- Any 8-K-equivalent filing naming the migration tool.
Gap: consensus estimates were searched and not found for this name at any fetched source. Recorded as unknown; no proxy was substituted.
9 · Synthesis & Verdict100%
A high-quality, wide-moat operator trading a little past its own base case. The quality score and the moat rating are earned by the filings; the valuation verdict is what keeps conviction at 61 rather than higher. The required margin of safety is not met at ₹1,563.20.
No imperative field exists in this schema. The dossier reports what the record supports; the decision and the order are yours, in your own exchange.
Rejected at file time
enforced in codeThese never reach a finished dossier. They are listed here because the check is the product.
- "operating margin improved 140bps year on year" → receipt #41unsupported — the figure does not appear in the stored text
- "management guided to 15% growth" → receipt #23unsupported — the stored text contains no guidance figure
212 receipts on file · 193 supported · 15 unverifiable · 4 unknown.
Assessment, not advice. Algoagent has no order endpoint — you decide and act in your own exchange.
IPO research reads registry and offer-document data. Algoagent cannot apply for an allotment, submit a bid, or move money.
Did the scores hold up, and is the doctrine answered. Reads only — nothing here writes or advises.
Calibration
| Conviction band | n | Mean | Median | Worst | Best |
|---|---|---|---|---|---|
| 80-100 | 6 | +18.4% | +16.1% | −4.2% | +41.0% |
| 60-79 | 11 | +9.2% | +7.8% | −14.6% | +33.5% |
| 40-59 | 9 | +2.1% | −0.4% | −21.8% | +19.7% |
| 20-39 | 4 | −6.7% | −8.1% | −28.4% | +9.2% |
MARGIN OF SAFETY MET
17 verdicts · mean +12.6% · median +10.4%
NOT MET
13 verdicts · mean +1.2% · median −0.9%
Top band separated from the bottom band by +25.1 points over the holding window.
Screens
64
Kills
19
Kill rate
29.7%
30 scored verdicts is above the 20 minimum, so percentages are emitted. Below it the surface says "insufficient sample" and prints nothing.
Appendix B checklist
ACME · 21/24 answered- #1Can you explain how the business makes money in two sentences?
- #2Is the moat source named, and is there evidence rather than assertion?
- #3Does ROIC exceed WACC, and for how many years?
- #7Has a 5% price raise been tested anywhere in the record?
- #11Is CFO tenure known?stage 5
- #14Is the pre-mortem written before the verdict?
- #18Is consensus known, and does the thesis differ from it?stage 8
- #21Is the ADR/GDR structure understood?N/A
- #24Is there a written kill criterion with a number in it?
Nine of twenty-four shown. An unanswered question names the stage that owes it.
MARKET WIRE
Market context, continuously.
Research context only — Algo Agent never executes or alters a trade.
Sources: Moneycontrol · Livemint · ET Markets · Business Standard
18 stories · 24h window
CATALYSTS
order win ×4capex ×3margin expansionRISKS
receivables ×2input costclient lossMIXED SIGNALS
guidance ×2pricingTop themes
- F
Industrial-automation order books widen for a third quarter, led by mid-cap suppliers
Moneycontrol
- F
Receivables cycles lengthen across the capital-goods complex as public-sector payments slip
Livemint
- R
Rupee steadies after the RBI's unscheduled dollar-sale window
ET Markets
Headlines · 7
9 stories · 48h window
Headlines · 2
0 stories · 72h window
No coverage in this window
The wire returned nothing for World risk in the last 72 hours. That is the honest empty state — no filler is generated to occupy the grid.
News informs research. It never places, sizes, or alters a trade.
Settings
Signed in
Keychain session
RM
r.mehta@example.com · Apple · signed in 12 Aug 2026 · later launches work offline
Algoagent stores your email, sign-in identifier and session timestamps. Not your research, holdings or provider keys.
Research profile
shapes the framework, not the verdictHorizon
3–5 years
Universe
Indian equity
Position cap
12% of book
Required MoS
20%
Risk tolerance
Moderate
Home currency
INR
Benchmark
Nifty 50 · standing
Reasoning model
your key, your bill- claude-opus-4ACTIVE
Anthropic · used for all nine stages
- gpt-5.1
OpenAI · signed in via device code
- gemini-3.5-flash
Google · demo assistant uses the site key — your Mac uses yours
Audit model
tools disabledRuns in a fresh context with no tools. It can object; it cannot re-score.
Broker credentials are read-only. Algoagent syncs holdings and never places an order — no order endpoint exists in the app.
Stored secrets
macOS Keychain · this Mac only-
AnthropicLLM
sk-ant-••••••••••••••••a3f9 · macOS Keychain · this Mac only
-
Zerodha KiteKITEREAD-ONLY
api_key ••••••••7c21 · access_token ••••••4d0e · macOS Keychain
-
IndianAPIFUNDAMENTALS
••••••••••••b117 · NSE/BSE fundamentals and price history
-
GenkaXBRLPAUSED
••••••••••••9e04 · Nifty 500 only · unlocks Piotroski, Altman, Beneish
-
TelemetryNONE
No usage data leaves the app, ever. There is nothing to store here.
no key
Reveal asks for Touch ID and re-hides the key after 20 seconds. Pausing parks a key without deleting it.
Holdings
read-only either wayRead-only sync · 8 holdings · last synced 09:41 IST · trade tokens are never requested
local · 0 holdings · zero-auth fallback
Zerodha expires Kite tokens at 6 AM IST. One click re-logs in and re-syncs; nothing else changes.
Algoagent has no order endpoint. Submitting, staging or prefilling a trade is structurally impossible, not merely disabled.
News providers
quota is local- Rss142/500358 left
keyless
- Newsapi38/10062 left
key set
- Gdelt11/∞no cap
keyless
Display
Follows the system by default. The toggle in this page's top bar drives the same setting.
⌘+ / ⌘− step the whole layout, so type stays crisp at each stop instead of scaling a fixed canvas.
Honoured from the system setting. Animations resolve to their final state rather than being killed mid-way.
Research
Six of nine stages. Required margin of safety rises 10 points and confidence is capped at 60%.
Evidence older than this is marked stale on the dossier.
Core
local processStatus
running
Uptime
4h 12m
Port
loopback only
Version
2.2.0
The diagnostics file deliberately excludes the core's tool log, which is where the name of every researched company and every fetched URL lives. Keys are reported as present or absent, never by value.
Help
Questions go to a person: contact@algoagent.in. Every release publishes a SHA-256 you can verify before opening the app.
Sample workspace · fictional book · research only, never advice. No key is held in this page. Read the product page · the full research method · risk disclosure