Features and data we have shipped recently, newest first.
Two new Metrics API types, app_review_count and app_rating, give the daily volume and average star rating of new Apple App Store reviews for tracked products with an app. Reviews no longer count toward mentions, which now measures news, social, video and forum conversation; review tone still counts toward sentiment and the SentiSense Score, and history was restated so charts read consistently across the change.
A new card at the top of the Analyst tab puts what covering firms say next to what social and news tone says, in the same present tense: the share of firms at Buy, Hold and Sell, the consensus target against the current price, and the SentiSense Score and Sentiment with their labels. The coverage endpoint now returns those firm counts as ratingBuckets, counted over the whole book so free and paid keys read the same numbers. Rating grades are stored in one spelling per rating, initiations never carry a prior grade, and a new analyst_reaction insight reports which firms moved targets or ratings in the five sessions after a company's print, with the denominator always stated.
Three new analyst endpoints and a rebuilt Analyst tab on every stock page. Ask who covers a ticker and get every firm active in the last twelve months, most recent first, with its latest price target, its current rating, the analysts it names and the notes that arrived without a byline; firms that rated a stock without publishing a target are listed too. Every named analyst now has a profile with the firms they have published under, their coverage book and their full call history, each note linked to the publisher's article. On the stock page the tab leads with a Rating Changes timeline, filterable by action, above the coverage roster, and has moved next to Financials. Free keys get the five most recently active firms on any stock; PRO gets the full book.
Every other earnings endpoint answers a question about one company. This one answers a question about the market: of everyone who reported in a given week, how many cleared the analyst estimate, and how the market actually traded them. The second half is the part nobody else publishes, and the two are joined in a single call. Ask for the last completed week, this week so far, the trailing year, or everything we hold. Each figure ships with its numerator, both of its denominators, the trailing baseline it is judged against, how complete the underlying data is, and the version of the rules that produced it, so you can re-derive every rate rather than trust it. Free with an API key.
The screener that powers the app is now a public API. Four endpoints let you filter roughly a thousand tracked US stocks, plus a separate ETF universe, on the SentiSense Score, attention, analyst consensus, technicals and price in a single call, with a field catalog to build from and 28 curated screens you can execute as-is. Free with an API key, and also available to agents through the MCP connector and the agent skill.
Market Mood now blends six signals instead of five: a new Options Flow constituent measures how broadly options positioning leans bullish or bearish across the tracked universe, smoothed over five sessions. The five existing social and price signals keep their relative proportions, and the API exposes the new signal as options_flow in the signal breakdown and optionsFlow in history rows.
The Earnings Analysis API is now documented. Its anchor endpoint diffs the Risk Factors section of consecutive 10-K and 10-Q filings and returns the passages a company added, removed, or rewrote, with a materiality score and a link to the filing. Free keys get the per-filing summary and totals; PRO unlocks the full diff detail.
The Earnings Analysis API gained a per-quarter analysis report: what a company reported, one object per fiscal quarter, carrying the headline, the KPI highlights that matter for that company with year-over-year deltas, the guidance language as management phrased it, and a summary of the earnings call. A companion feed answers the cross-ticker question of who reported in the last week, so you can sweep the covered set and then pull each quarter. Free keys get the latest quarter plus the full recent feed, and PRO unlocks every stored quarter. The Python and Node SDKs ship both as well.
stock-earnings-analysis reads a quarter the way an analyst does: the analysis of what was reported, the risk-factor diffs attached to the quarter they belong to, who reported in the last week, and the forward calendar. last-30-days-in-markets synthesizes the month into one brief, covering the day-by-day arc of market mood, the biggest story themes, and the signals that accumulated, with every claim carrying its date and its real coverage window. Both install with the rest of the collection: npx skills add SentiSenseApp/skills.
The Indexes API is now documented. Market Mood, Fed Sentiment, and AI Sentiment all answer on the same envelope, so one integration covers every current and future SentiSense index, and Market Mood now resolves here too alongside its fuller view at its own endpoint. The Python and Node SDKs ship it as well, with methods to list the indexes, read one, and pull its history.
The Insights API reference used to name ten of the types it can return, four of them under keys that no longer matched the response. It now lists all forty-eight types the API can serve, each with the key you filter on and what the signal means, so you can build against the full set instead of discovering types at runtime. A test keeps the page and the served types in step.
Three reads are now documented as their own endpoints: a one-call sentiment headline for any covered ticker, the eleven GICS sectors measured against the market's own tone, and a breadth line showing the bullish, neutral, and bearish share of coverage. Previously these took several calls to assemble.
Sentiment is no longer limited to stock symbols. CEOs, politicians, products, and institutions each carry their own tracked entity, so you can follow the tone around a person or a product line as its own series.
The connector picked up a stock screener, an earnings calendar lookup, and a reported-financials tool. An assistant can now go from a screen, to a company's statements, to the date of its next print without leaving the conversation.
A full options vertical: IV rank and percentile, unusual activity, put/call positioning, and the walls that matter for a given expiry, across roughly a thousand tickers. Available in the app and over the API.
Finance YouTube is now ingested and scored alongside news, Reddit, X, and Substack. It shows up in the per-source breakdown, so you can see when a move in tone is coming from video rather than from the wires.
Every major institutional filer now has its own page: portfolio size, position count, and what it bought and sold last quarter, including parent and subsidiary rollups. Start from a fund and work down to the position, instead of starting from a ticker.
The Financials tab now renders complete income statement, balance sheet, and cash flow tables with a timeline you can scrub back through prior periods. The same history is available over the API for multi-period trend work.
SentiSense is now available as a hosted MCP connector, so an assistant that speaks MCP can pull sentiment, filings, and market data directly into a conversation. No SDK, no glue code: connect it once and the tools show up.
A daily index measuring tone across the companies driving the AI cycle, so the theme can be tracked as one thing rather than ticker by ticker. It sits alongside Market Mood and Fed Sentiment on the indexes surface.
Upcoming earnings dates are now a first-class API surface, with a discovery endpoint that lists the calendars we publish. Agents can schedule research around a print instead of scraping a date from a news page.
PRO API keys no longer have a monthly request ceiling. Usage is now bounded only by a per-minute rate limit, so a daily-driver integration or an always-on agent can run without budgeting calls against a monthly quota.
A tracker for the tickers retail traders are actually discussing, ranked by mention volume and tone rather than by a single loud thread. It updates continuously, so a name climbing the list is a live signal and not a weekly recap.
A weekly index that tracks how the market is talking about Federal Reserve policy, built from the same sentiment pipeline that powers per-stock coverage. Useful for reading the mood into an FOMC week rather than after it.