AI Stock Analysis Grounded in Real Fundamentals
Ask any equity question and get a concise, cited research read built from live fundamentals — not hallucinated numbers from a model's memory.
US coverage included in your free 14-day trial · no credit card required
StoqPulse AI is a research copilot that answers from facts, not vibes. Before it replies, the server pulls a live context block — quote, financial statements, ratios, fair value, technicals, analyst targets, ownership, earnings, filings, and news — for the ticker you asked about, and the model is instructed to use only those figures. US coverage and the default model are free for 14 days (no card), so you can pressure-test a thesis before choosing a paid plan.
Answers built from live data, not model memory
When you ask about a ticker, StoqPulse assembles a live stock-analysis context block from the same fact set behind your Analysis tabs: quote, fundamentals, financial statements, revenue segments (product and geographic mix with a multi-year trend), valuation ratios, fair value (DCF), peer comparison, quality/risk scores, technicals, analyst consensus, dividends, ownership, earnings, transcripts, SEC filings, and recent news. The system prompt forbids using remembered numbers for any financial figure — revenue, EPS, margins, growth, price, market cap. If a fact isn't in the live block, the AI says so rather than inventing one.
Every figure cited to its fiscal period
The model is required to report the as-of date or fiscal period for any financial figure — FY2025, Q3 2025 — and to use the most recent period present in the live context. If only an older period is available, it states that period explicitly instead of implying it's current. Ask 'why did NVDA's F-Score drop' or 'compare MSFT and GOOGL on quality factors' and you get the metric and its implication in the same line, with the period attached, so you can check the claim against the source tab.
Grounded in your own portfolio, watchlist, and journal
For signed-in users, the server adds an authoritative account context block straight from your data: portfolio holdings and weights, watchlist names and price targets, your composite-score factor weights, and recent decision-journal theses. The AI uses it to ground concentration, portfolio-watchlist overlap, and thesis-versus-news context — and it leads with the factors you weight most heavily. It will not invent holdings, targets, or journal entries that aren't in that block.
Research, not recommendations
StoqPulse AI is built as a research assistant, not an advisor. It never tells you to buy, sell, hold, trim, or size a position, and it never predicts prices as expectations. Ask 'should I buy X' and it pivots to factual analysis — fundamentals, valuation, technicals, ownership, filings, transcripts, news, historical performance. Backtest and simulation output is always framed as hypothetical and historical. A short disclaimer accompanies any security-specific answer. The boundary is deliberate: honest research you can act on yourself.
In-app help that knows the actual app
Ask 'where do I set a price alert' or 'how do I log a decision' and the AI answers from a built-in application guide describing real screens, sidebar navigation, tabs, and button labels — not from generic guesses. It names the exact screen and the real control, and skips the investment disclaimer for pure walkthroughs. If something isn't covered in the guide, it says so and points you to the closest relevant screen instead of fabricating steps.
Your default model, or bring your own key
The default AI model needs no setup and no key — ready to use immediately. Prefer your own provider? Bring your own key for supported third-party AI providers; keys are resolved only inside the AI service and the browser never forwards raw provider credentials. After each answer, in-scope follow-up chips appear — research-only questions scoped to names you own or watch, or macro topics — so you can keep digging without retyping context.
Frequently asked questions
Before answering, the server fetches a live context block of real fundamentals for your ticker and instructs the model to use only those figures — never remembered numbers. Every financial value must carry its fiscal period, and if a fact is missing from the live context the AI says so instead of guessing. It's the same data shown across your Analysis tabs.
US coverage and the default AI model are free for 14 days with no card, so you can ask grounded questions before choosing a paid plan. Monthly AI-query limits are plan-driven and read from live pricing config. See live pricing for the exact allowance on each tier.
No. StoqPulse AI is a research assistant by design — it never recommends buying, selling, holding, trimming, or sizing a position, and never predicts prices as expectations. Ask an advice question and it pivots to factual analysis: fundamentals, valuation, technicals, ownership, filings, and news, with a disclaimer on security-specific answers.
For signed-in users, yes. The server adds an authoritative context block from your holdings, watchlist, price targets, scoring weights, and recent decision-journal theses. The AI uses it to ground concentration, overlap, and thesis-versus-news analysis, and leads with the factors you weight most. It won't invent holdings or entries you don't have.
Yes. The default needs no setup, but you can bring your own key for supported third-party AI providers. Keys are resolved only inside the AI service — server-side from your encrypted settings or platform secrets — and the browser never reads or forwards raw provider credentials.
The same fact set as your Analysis tabs: quote, fundamentals, financial statements, revenue segments (product and geographic mix), valuation ratios, fair value, peer comparison, quality and risk scores, technicals, analyst consensus, dividends, ownership, earnings and transcripts, SEC filings, and recent company and event news — assembled live per question.
Explore more StoqPulse tools
Start free for 14 days — US coverage included
Build the discipline into your process. No credit card required.