PRIVATE PILOT · Provided free of charge for evaluation and feedback only · No commercial relationship is created by use of this platform
HAMSALENS / news-sentiment

Market & News-Sentiment Terminal · 🇺🇸 US & 🇮🇳 India

Every breaking headline, scored, mapped to a ticker, and tracked for what happened next.

HamsaLens is a professional-grade market dashboard built around one idea: news can move markets before prices fully reflect it. It reads incoming news continuously, scores each story's sentiment with AI, maps it to the stocks it references, and keeps a transparent record of whether the market moved in the direction the signal indicated — across ten connected modules, now covering two markets: US equities and India (NSE), with optional catalyst alerts delivered to Telegram.

Get the alerts on Telegram · free to read, no account needed

Every catalyst signal is posted to a public Telegram channel the moment it is generated, before the outcome is known — then the result is posted as a reply once the window closes. Open either channel and watch it happen live. See how the record works ↓

The walkthrough

Ten modules, one connected view of the market

Each module answers a different question a portfolio manager asks during the day — from "what just broke?" to "is this desk's model actually any good?" Here is every one, in order.

01

🌐 Live Terminal News sentiment

The core screen. Every incoming headline is classified bullish, bearish or neutral, scored from 0 to 10, and mapped to the tickers it references — with its latest available price beside each one. The engine then observes the market and marks each signal ✅ or ❌ depending on whether the stock moved in the direction the signal indicated, so the feed becomes a transparent, self-scoring record rather than a wall of headlines. A "News-Sentiment Leaders" panel ranks the day's most-discussed names by net sentiment.

02

⭐ Watchlist Your book

The same sentiment, price and signal treatment, narrowed to the tickers you actually hold or follow. Add your positions once and the terminal filters every incoming story down to what touches your book — so a headline about a name you own never scrolls past unseen.

03

📈 ML Performance Matrix Transparent track record

Most tools hide how often they were right. This one shows it. The matrix reports the historical hit rate and total signals recorded, broken out across multiple time horizons — how each signal looked 90 minutes after the news, by the same-day close, and by the next-day close. It is a deliberately transparent scoreboard of past results, and it updates as more signals are recorded. Past results do not guarantee future outcomes.

04

🔥 Sector Heatmap Rotation at a glance

Aggregates every scored story up to the sector level, so you can see where sentiment is concentrated — which sectors the news is bullish on, which are under pressure, and how one-sided the flow is. Click any sector to drill into the individual names driving it.

05

📊 Global Indices World tape

The major benchmarks on one screen — S&P 500, Nasdaq, and international indices including the NIFTY 50, CAC 40 and Euro Stoxx 50 — with levels and moves, so a US desk keeps one eye on how Europe and Asia are setting the tone.

06

💱 Currencies FX & crypto

Major currency pairs and leading crypto assets, tracked alongside the equity view — the macro backdrop that frames every single-stock call.

07

🏆 Commodities Macro drivers

Gold, oil and the commodity complex, so the moves that ripple into energy, materials and inflation-sensitive names are never off-screen when you are weighing a sector call.

08

⚡ Options Flow Activity view

Displays notable options activity — where trading volume and open interest are concentrated across calls and puts — as informational market context only. It is not a forecast, and does not predict future moves in the underlying.

09

🏛 Index Members Constituents

Expand any major index into its full membership with live prices and daily moves — the fast way to see whether an index move is broad-based or carried by a handful of names.

10

📅 Earnings Radar Estimates vs actuals

A forward calendar of who reports and when — before the open or after the close — with consensus EPS and revenue estimates, the actuals the moment they land, and an automatic beat / miss verdict. Filter by sector, and choose how wide to cast the net, from your tracked names to any large-cap in the news that week.

Coverage

Two markets, one terminal

The same ten modules and the same sentiment engine now run against two separate markets, each with its own news sources, trading calendar, currency and sector taxonomy. You pick a market at launch — the terminals are fully isolated, so US and India data never mix.

🇺🇸

US Terminal NYSE & Nasdaq

The original terminal. Headlines are drawn continuously from a broad set of publicly available US news and corporate-disclosure feeds — including newswire releases and regulatory filings — then scored, mapped to US tickers and tracked against the 9:30–16:00 ET session. Prices, indices, options-activity and earnings modules are all tuned to US market structure, quoted in US dollars.

🇮🇳

India Terminal NSE · New

A dedicated India build, not a translation of the US one. It reads Indian financial media and exchange announcements, follows the 09:15–15:30 IST session and NSE holiday calendar, and prices everything in rupees (₹), with large figures shown in crore and lakh. Sectors follow an Indian taxonomy — RBI/SEBI policy, NBFC & financial services, IT services, auto ancillaries, FMCG — and the Earnings Radar tracks the NIFTY 50. Global index tiles add Asian, European and US benchmarks for context.

Exchange-validated tickers Accuracy

Company names in Indian headlines are notoriously ambiguous, and language models resolve them unreliably. Every symbol the model proposes is therefore checked against the live NSE equity list before it is shown, and unrecognised symbols are discarded rather than guessed — so a story is either mapped to a real listed company or left unmapped.

Under the hood

A sentiment engine that grades its own homework

The difference between a news feed and a signal is accountability. Every call the terminal makes is logged, checked against the tape, and fed back into how much the next call is trusted.

STEP 01

Read & score

Headlines are sourced from publicly available news feeds, classified for sentiment and scored 0–10 by an AI model, then mapped to the tickers and sector each story references.

STEP 02

Weight by track record

A raw score becomes an adjusted score by re-weighting it against how that sector-and-direction signal has behaved across the terminal's own recorded history — so historically consistent patterns are weighted more heavily than those with mixed track records.

STEP 03

Check & learn

Once the session has settled, each signal is marked ✅ or ❌ and written to the record. That record feeds the ML Performance Matrix and informs how future signals are weighted, so the scoring reflects more history over time.

STEP 04

Measure honestly

How you score the scoreboard matters as much as the scoreboard. Three rules keep the record from flattering itself: each story is classified by event type and only genuine company catalysts — results, deals, guidance changes, regulatory actions, contract wins, management changes — are graded, while opinion columns and multi-stock round-ups are shown but excluded; the same story echoed by several outlets is counted once, at first publication; and a move is measured net of the benchmark index over the same window, so a broad market rally is not recorded as the signal being right.

Measurement methodology, applied to both markets. Recorded accuracy is published inside the terminal's ML Performance Matrix, including the periods where it is unremarkable. Past results do not guarantee future outcomes.

Open record

Don't take our word for it. Watch it happen.

Measuring yourself honestly is easy to claim and hard to prove. So every catalyst signal is also published to a public Telegram channel the moment it is generated, before the outcome is known — then the result is posted as a reply to that same message once the window closes. Winners, losers and everything in between. No account needed to read it.

Posted before the outcome exists Forward-locked

The signal goes out first and the result follows as a reply to it, so the original text and its timestamp cannot be adjusted afterwards. The entry price used is the first trade at or after the message was published — never a price from before the news — so only a move that was actually available is ever claimed.

Graded net of the index Market-adjusted

Each result is measured against the S&P 500 or NIFTY 50 over the identical window, so a broad rally is never recorded as the signal being right. Outcomes inside a ±1% band are published as flat rather than quietly dropped, and the holding window never crosses overnight.

Nothing is removed Sequential

Channel posts are numbered in sequence, so a deleted message would leave a visible gap that any reader can check. The inclusion rule is fixed, published in each channel, and applied before any outcome is known — a signal cannot be selected in or out on the strength of how it turned out.

Two tiers, both shown CORE & WATCH

Higher-conviction catalysts are tagged CORE and lower-conviction ones WATCH, and both are published. Keeping the weaker tier visible is the point: it is what makes the stronger tier's record meaningful rather than self-selected.

These channels opened on 4 August 2026 and fill as the sessions run, so the record is deliberately young — it is an open log being built in public, not a performance claim. Signals are informational research output, not advice or a recommendation to trade. Past results do not guarantee future outcomes.

Off-screen

The high-conviction ones find you

Nobody watches a screen all day. The terminal can push its strongest catalysts straight to a private Telegram channel — one per market, so an Indian session never wakes you for a US headline and vice versa.

Filtered, not firehosed Signal over volume

An unfiltered "score 6 and above" rule was measured at several hundred messages a day — unreadable within a week. Alerts therefore apply the same discipline as the scoreboard: a story must carry a high conviction score, name a real listed company, be a genuine catalyst rather than commentary, and not repeat a story already sent. What arrives is a short, readable stream rather than a feed you mute.

What a message looks like At a glance

Direction and score, the companies referenced, the headline, the event type and sector, the model's one-line reasoning, and a link back to the source article — enough to judge in a couple of seconds whether it deserves your attention, without opening anything.

Tuned to you Pilot

The conviction threshold is a setting, not a rebuild — it can be raised for a quieter channel or lowered for fuller coverage, per market. Alerts are informational notifications about scored news, never instructions to trade. Available on request during the pilot; delivery depends on third-party messaging infrastructure and is not guaranteed.

Open the terminal

The full dashboard runs in your browser — no install, no terminal hardware. Step inside and watch incoming news headlines get scored and mapped as they arrive.

Important information

Disclosures & risk disclaimer Read