Research

Helper tools that round out the pre-trade flow with source-cited news and sentiment research. The tools return operating instructions — the LLM does the web research.

crowdtrendz_crypto_news_research

Kicks off a structured news & sentiment report for a given token. The tool itself returns operating instructions for the LLM — the LLM is then expected to use its own web-search capability to gather sources, and to render a source-cited markdown report directly in chat.

Server-side or model-side?

The tool does not fetch news server-side. The model does the research with whatever browsing tool it has (claude.ai's web search, ChatGPT browsing, etc.) and inlines the report. CrowdTime supplies the rubric.

Arguments:

Name	Type	Default	Notes
symbol	string	—	Token symbol or name (BTC, ETH, SOL, …)
horizon	string?	immediate + short-term	Forward-looking analysis window (next 7 days, Q2 2026, …)
lookback	string?	model judgment	Backward recency cap on sources (last 7 days, since 2026-04-22, …)

The model will typically ask you for horizon and lookback before running, unless you say "use defaults" or specify them inline.

Output shape

The model produces a markdown report following a strict rubric: executive summary, key developments, sentiment analysis with a High/Medium/Low confidence label, potential catalysts (split by horizon), outlook with an actionable takeaway, and a numbered Sources section with inline citations like [1][2] throughout.

Example prompt:

Run the crowdtrendz_crypto_news_research on ADA, horizon next 14 days, lookback last 10 days.

crowdtrendz_stock_news_research

Same pattern as crowdtrendz_crypto_news_research, but scoped to equities and ETFs. The tool returns operating instructions for the LLM to do its own web research and inline a source-cited markdown report. Sourcing and analysis rubric are adapted for stocks: primary filings (10-K / 10-Q / 8-K, RNS announcements), reputable financial press (Reuters, Bloomberg, FT, WSJ, CNBC), analyst actions, insider transactions, and benchmark / sector correlation.

Arguments:

Name	Type	Default	Notes
symbol	string	—	Stock or ETF ticker (NVDA, MSFT, VUAG, BP, …). Add an exchange suffix (e.g. BP.L) only if the ticker is ambiguous.
horizon	string?	immediate + short-term	Forward-looking analysis window (next 7 days, next earnings, Q2 2026, …)
lookback	string?	model judgment	Backward recency cap on sources (last 7 days, since 2026-04-22, …)

The model will typically ask you for horizon and lookback before running, unless you say "use defaults" or specify them inline.

Output shape

Same structure as the crypto report — executive summary, key developments, sentiment analysis with a High/Medium/Low confidence label, potential catalysts (split by horizon), outlook with an actionable takeaway, and a numbered Sources section with inline citations like [1][2] throughout. The outlook also flags the next known dated catalyst (earnings, ex-dividend, scheduled regulatory decision) when one falls inside the horizon.

Example prompt:

Run the crowdtrendz_stock_news_research on NVDA, horizon into next earnings, lookback last 14 days.

crowdtrendz_stock_dividend_research

Same pattern as the news agents, but scoped to dividend events: declarations, declared-amount changes (increases / cuts / suspensions / specials), upcoming and recent ex-dividend dates, record dates, and payment dates. Always anchored on an exchange / region / index (NASDAQ, NYSE, LSE, FTSE 100, S&P 500, all-US, …), and optionally narrowed to a single ticker (AAPL, BP) within that scope. The tool returns operating instructions for the LLM to do its own web research and inline a source-cited markdown report.

Sourcing prioritises issuer-primary disclosures (press releases, SEC 8-Ks, RNS announcements, exchange dividend calendars) over aggregators, and the rubric requires cross-checking every dated or numeric claim against at least two sources.

Arguments:

Name	Type	Default	Notes
exchange	string	—	Required. Exchange / region / index scope that anchors the report (NASDAQ, NYSE, LSE, FTSE 100, S&P 500, all-US, …). When symbol is also set, provides peer / sector context for that ticker.
symbol	string?	—	Optional single ticker to narrow the report within exchange (AAPL, BP, …). Add an exchange suffix (e.g. BP.L) only if ambiguous.
event_types	string?	all	Comma-separated filter: declaration, change, ex_dividend, record, payment.
horizon	string?	next 14 days	Forward window for upcoming ex / record / payment dates (next 7 days, this month, Q3 2026, …).
lookback	string?	last 14 days (model judgment)	Backward recency cap for declarations and declared-amount changes (last 7 days, since 2026-04-22, …).

exchange is always required; symbol is optional and narrows the scope to a single ticker within that exchange. The model will typically ask you for horizon and lookback before running, unless you say "use defaults" or specify them inline.

Output shape

A markdown report with:

• Executive summary (2-3 sentences) with a High / Medium / Low data- quality confidence label
• Declared dividend changes (grouped into Increases, Cuts / Suspensions, Specials)
• Upcoming ex-dividend dates (grouped chronologically: This Week, Next Week, …)
• Upcoming payment dates
• Brief sustainability note per issuer (payout ratio, FCF cover)
• A numbered Sources section with inline citations like [1][2] throughout

Each event bullet bolds the ticker and inlines: declared amount, ex date, record date, payment date, indicative forward yield, and change vs. the prior dividend (absolute and %).

Example prompts:

Run the crowdtrendz_stock_dividend_research across NASDAQ, horizon next 7 days, lookback last 14 days.

Run the crowdtrendz_stock_dividend_research on BP across LSE, horizon next 30 days, use defaults for lookback.

Run the crowdtrendz_stock_dividend_research across the FTSE 100, event_types change,ex_dividend, horizon this month.