Every development in the Intelligence table now shows the story it belongs to, so you can follow how news about NVIDIA, Amazon or SpaceX unfolds without leaving the table.
Added a Story column showing where each development sits in its story, such as 4 of 7, with brand-new developments marked New.
Opened every row onto the story so far: the developments around it, the story and topic it belongs to, and links to both.
Added Trending and Latest views to the Intelligence page, with Trending listing the newest development of each story people are talking about most.
Listed each topic’s stories in a full-width table on its page, ranked the way the topic ranks them.
Let the topics list scroll back through every topic, newest update first, and removed the default Influence and Credibility filters so every development shows.
The table on Home is now a screener: every asset in the selected market, ranked by attention, with its price, mindshare, news, conversation, communities, sentiment and mindshare forecast on one row.
Replaced the intelligence table on Home with a screener that follows the Equities, Commodities, Indices and Crypto switch and ranks every asset by mindshare, with the intelligence feed still on the Intelligence page.
Drew each asset’s last day of price and of mindshare beside its values, with the 24-hour moves alongside.
Added a Mindshare forecast column: the next day’s forecast fan for each asset, the share it lands on, and the forecaster’s call when it is confident.
Counted the intelligence stories that began about each asset in the last 24 hours, hour by hour, next to its posts and their hourly sentiment.
Showed which communities are talking about each asset and the bullish, neutral and bearish split of its posts, with every column sortable.
The new Home ranks the stories the world is talking about right now, from NVIDIA and Amazon to SpaceX, politics and sport, and shows how each one developed.
Led Home with the top five stories in rotation: each with its timeline, its last 24 hours of conversation coloured by sentiment, who is talking and its sentiment split.
Ranked stories by the conversation they have now and how important they are, so the latest important news rises to the top and yesterday’s fades.
Measured each story by a filter built from the tags on its own posts, in every language, so a story’s chart shows the conversation about that event rather than everything said about its subjects.
Showed whether each story’s conversation is speeding up or slowing down, and grouped follow-up stories into the story they develop.
Added Chat about this and Dashboard to the lead story, to ask Nebula about it or have it build a dashboard for it.
Moved the market views (Market Digest, Mindshare, Sentiment Forecast and Catalyst Calendar with the intelligence table) to the Intelligence page.
Constellation 1 now supports binary, multiple-choice and numerical questions alongside event discovery.
Added explicit question formats with resolution rules, deadlines, answer options and numerical units to the forecast API.
Reused one research brief across the ensemble and saved individual answers, the final forecast and recorded costs.
Routed question forecasts through Constellation without additional simulation or curation calls.
Extended evaluation results to show option probabilities, numerical distributions and matched external scores, with retrospective limitations made explicit.
The Nebuliser is now Orion, a live graph that combines what is discussed alongside a stock with the relationships people claim it has.
Renamed the Nebuliser to Orion, with existing links redirecting to the new page.
Rebuilt the graph as a live network you can drag, pan and zoom, and built it from several stocks or subjects at once.
Added a Relationships list that reads each relationship as a short statement, such as a former CEO or a new partnership, with its confidence, claims and authors.
Combined the Conversation table and the Relationships list across every subject in the graph, leading with new relationships and those between your chosen subjects.
Filtered relationships by one or more kinds of subject and by confidence, with low-confidence relationships hidden by default.
Added a Relationships tile to stock and topic pages that lists what is said about the subject and opens Orion when expanded.
Every insight in Intelligence now shows how big it is and how stocks like NVIDIA and Apple reacted, forecast when the news breaks and replaced by what actually happened as the day plays out.
Added an Impact score from 0 to 100 to each insight, coloured bullish, bearish or neutral, with a forecast mark until the reaction has been measured.
Made rows open in place: an opened insight shows its summary, the conversation around it and the clusters talking about it, without leaving the table.
Added a breakdown for every subject an insight names: attention compared with its usual level, the shift in sentiment, and the forecast likely range before the result is in.
Added price reactions for stocks and other priced assets: the price across the 24 hours either side of the insight, the move at 1, 6, 12 and 24 hours, and the live move for the first two days.
Added Sentiment and Price columns to the widest Intelligence table, and removed the Tags and Type columns.
Intelligence now writes up developments as they surface instead of every ten minutes from the most-discussed names, so news about stocks like NVIDIA and Amazon, and about the smaller companies around them, reaches you within minutes and keeps up as the story moves.
Changed Intelligence to read every incoming post and news article as it arrives, so developments on less-discussed stocks, sectors and markets are covered, not only the day's most-discussed names.
Made credible reports appear within minutes; a claim from ordinary accounts is held briefly for corroboration, then written as an attributed claim.
Made stories update in place when corrections or new facts arrive, while a later development, such as a deal closing after it was announced, became its own story.
Labelled rumours by how the claim is sourced, such as a report citing people familiar with the matter, and relabelled them when the company or an official confirmed or denied them.
Added coverage of remarks by market-moving figures, from central bankers to prominent investors, and of significant market moves and technical analysis that names its levels.
Home now shows where the conversation is heading for the names that matter, so you can see at a glance whether sentiment on stocks like NVIDIA and Tesla is set to build or fade over the next one to three days.
Added a Sentiment Forecast tile to Home: three stocks side by side, each with its recent sentiment and the forecast branching out from now, its likely range green above zero and red below.
Added a list to choose the names: Trending shows the most discussed; Bullish and Bearish show well-discussed names the forecast is confident will end higher or lower.
Added a +24H or +72H horizon, with the column showing where sentiment is likely to land at the point you hover.
Added hover readouts with the date, median and likely range at any point, and a full-screen view with the full landing column.
Made forecast fans easier to read on every chart: simulated paths no longer brighten where they cross, and the forecast hatching is lighter.
The 72-hour Sentiment forecast behind the Nebula chart is now available over the API, MCP and CLI, so you can pull where bullish and bearish conversation about stocks like NVIDIA is likely to go into your own models and alerts.
Added GET /forecasts/sentiment to the Forecasts section: every 4 hours, the bullish, bearish, neutral and total posts expected in the step, each with its percentiles, and the sentiment signal with percentiles and 21 quantiles, plus 24 sample paths.
Added the current sentiment signal, whether sentiment is turning bullish or bearish, and the chance the signal is higher in one and three days, each with a flag for when that chance leans clearly one way.
Made GET /forecasts/assets list coverage for both forecasts, since every asset with a Mindshare forecast has a Sentiment forecast too.
Priced it like the Mindshare forecast: 5 credits, 10 with refresh=true, and free for an asset that cannot be forecast, up to 30 an hour per account; errors, including a forecaster outage, cost nothing.
The Sentiment chart on stock pages can now show where the crowd’s mood is likely to go over the next three days, so you can see whether bullish or bearish conversation about stocks like NVIDIA and Tesla is likely to build or fade.
Added a forecast toggle to the Sentiment chart that draws a fan forward of the line: simulated paths, a dashed median path and a likely range covering the middle 80% of outcomes, green above zero and red below.
Added forecast post-volume bars beside the history, showing the likely number of posts in each four-hour step and how high it could reasonably go.
Added a column beside the chart showing where sentiment is likely to land, and a header readout with its likely range; the chance it ends higher is shown only when the forecast is confident.
Made the forecast follow a shift in mood as it happens: when sentiment has just turned, the fan reflects how similar turns have played out.
Added the forecast to the Overview price chart: its Sentiment and Mindshare panes open with the forecast and the upcoming events due in the next three days, with a switch in each pane’s header and the pane’s readout following the forecast as you hover it.
Changed both the Sentiment and Mindshare forecasts to use the median as the centre line and to write ranges as “low to high”.
The 72-hour Mindshare forecast behind the Nebula chart is now available over the API, MCP and CLI, so you can pull where attention on stocks like NVIDIA is likely to go into your own models and alerts.
Added GET /forecasts/mindshare in a new Forecasts section: the forecast distribution every 4 hours with percentiles and 21 quantiles, 24 sample paths, the chance Mindshare is higher in three days, its likely range and the chance of a sudden surge.
Added GET /forecasts/assets, the list of every asset with a forecast right now, so you can check coverage before asking for one.
Added refresh=true to re-run the forecast on the latest data instead of reading the most recent five-minute run.
Made an asset that cannot be forecast a successful answer with status not_forecastable and the reason, such as too little history or no recent price.
Priced forecasts at 5 credits and 10 with refresh=true; an answer for an asset that cannot be forecast is free, up to 30 an hour per account, and errors, including a forecaster outage, cost nothing.
Social Momentum can now show which author clusters lead the conversation about a stock like NVIDIA, and who is climbing or falling, instead of only how much each posts.
Added a rankings toggle to the Social Momentum chart that ranks every cluster by unique posters in each slice of the selected window.
Showed each cluster’s climb or fall over the period with green and red arrows and the number of places moved, beside its icon and name.
Kept the leaders in view and let the remaining clusters scroll inside the tile, with a count of how many sit below.
Added hover detail for any cluster and slice (rank, unique posters and change from the previous slice), with click to lock a cluster.
The Mindshare chart on stock pages can now show where attention is likely to go over the next three days, so you can see whether conversation about stocks like NVIDIA and Tesla is likely to build or fade.
Added a forecast toggle to the Mindshare chart that draws a fan forward of the line: simulated paths, a dashed average path and a dotted likely range covering the middle 80% of outcomes.
Added a header readout with the chance Mindshare is higher in three days and its likely range, and a column beside the chart showing where it is likely to land.
Added hover detail for any point in the next three days: the average path, likely range and chance above today’s level.
Built the forecast from the latest 15 minutes of posts, author-cluster activity and price moves, refreshed about every five minutes.
Made the toggle explain why when a forecast is unavailable, such as too little recent conversation, no recent price, or an author-cluster filter being selected.
The clusters page now lists every cluster in one table and compares them side by side, so you can see which groups are early and where they disagree on stocks like NVIDIA and Tesla.
Replaced the cluster cards with a table of every cluster: daily attention bars, posts and their change, stance against the market, the name each group turned to more than the market this week, and its bullish/bearish split, with each row opening that cluster’s page.
Added Profile Extremes, showing the highest and lowest cluster on every profile trait.
Added Recent Hype Cycles, showing the crowd’s attention hour by hour around each spike and the order the clusters arrived in, beside Attention Timing for every cluster with its likely range.
Added Most Divisive Assets, the widely discussed names the clusters are most evenly split on.
Added the cluster pages’ asset-class selector to the clusters page, scoping every view to any mix of equities, commodities, indices and crypto.
Four new endpoints return the clusters page’s comparisons across every author group at once, for equities, commodities, indices and crypto, through REST, MCP and the CLI.
Added GET /clusters/comparison: every cluster’s posts and their change, its bullish share against the market’s, the asset it turned to more than the market did this week, and its daily posts.
Added GET /clusters/comparison/traits: every behavioural trait across every cluster, highest first.
Added GET /clusters/comparison/hype-cycles: the biggest hype cycles, the market’s posts hour by hour around each peak, and the order the clusters arrived in.
Added GET /clusters/comparison/divisive-assets: the widely discussed assets the clusters are most evenly split on, with every cluster’s stance.
Meridian brings three prediction streams into one curated view of potential stock and global-market events.
Added MetaForecaster to run Superforecaster alongside Simulac, feed the saved simulation to Simulation-informed Forecaster, and curate up to twelve events by default.
Separated event probability from confidence in the estimate, with explanations and references to the underlying prediction streams.
Saved the stream artifacts and final forecast so completed requests can be revisited without repeating the model work.
Every cluster page now shows how that group behaves across stocks, commodities, indices and crypto: how early it moves, where it disagrees with the market, and where its attention is rotating.
Added a Profile that ranks the cluster against every other on earliness, contrarianism, hype fade, conviction, focus and breadth.
Added Attention Timing, Stance vs Crowd, Rotation and Cluster Overlap tiles, with the screener alongside them.
Added an asset-class selector to the cluster search bar that combines equities, commodities, indices and crypto, or narrows to any mix of them, across every tile on the page.
Replaced the cluster page’s call cards with the combined Calls tile from asset pages, and its summary bars with the headline stats row.
Five new endpoints show how any author cluster behaves across stocks, commodities, indices and crypto: how early it is, where it disagrees with the market, and where its attention is moving.
Added GET /clusters/profile, which ranks one cluster against every other on earliness, contrarianism, hype fade, conviction, focus, breadth, bullishness, discovery, turnover and coverage, with the full peer distribution for each trait.
Added GET /clusters/hype-cycles, which compares how a cluster’s attention builds up against the market’s around every market attention peak and reports how many days ahead of the market it runs.
Added GET /clusters/stances, which sets a cluster’s net stance on the stocks and other assets it discusses most beside the market’s, with its contrarian share against its peers’.
Added GET /clusters/rotation and GET /clusters/overlap, which show the assets gaining and losing share of a cluster’s attention, its weekly asset-class mix, and which other clusters watch the same assets and whether they lean the same way on them relative to the market.
Let every cluster analytics endpoint combine equities, commodities, indices and crypto in one asset_class selection, and return each asset as the asset_id and asset_class pair the other endpoints accept.
Two new author clusters ranked by timing rather than returns show who consistently starts talking about a stock before its attention takes off, and who consistently shows up once it has already peaked.
Added Early Movers, the accounts most consistently among the first to talk about an asset as its attention begins to spike, rebuilt daily from their last 90 days of spikes.
Added Late Movers, the accounts most consistently arriving at the peak of an asset’s attention or as it fades; when they pile in, attention has usually run its course.
Made both available everywhere Smart Money is: the sentiment, mindshare and social momentum cluster selectors, cluster pages and the home clusters row, calls and Discovery rankings.
Matched each post to the lists as they stood on the day it was written, so a chart of what Early Movers said about NVIDIA last month reflects who was on the list then.
Exposed early_movers and late_movers through the clusters parameter of the Developer API, MCP and CLI.
A new chart on every stock and token page shows whether the conversation about an asset is running ahead of, or behind, the money actually trading it.
Added the Social vs Trading Volume Index to the Performance view of stock and token pages, drawn against the price.
Compared a trailing week of mentions with a trailing week of dollar trading volume, each measured against the asset’s own 90-day normal, on a 0–100 scale where 50 means the two are in their usual proportion.
Added a bars toggle that draws each day’s trading volume rising from the bottom and posts hanging from the top, each on its own scale, behind the line.
Drew the line rising in blue as trading runs ahead of the conversation and falling in pink as the conversation runs ahead of trading.
Made the index available as a custom dashboard tile.
Added GET /social-volume-index to the Developer API, MCP and CLI, returning the daily index with both of its components and the day’s raw posts and dollar volume.
Requests that end in an error are no longer charged, plan limits say how to lift them, sentiment responses stop repeating their main series, and search ranks topics above low-value tokens that share their name.
Refunded every fixed-price request that ends in an error, such as an unknown asset or a server failure, and reported the refund as X-Credits-Charged: 0.
Fixed rejected parameters returning a second JSON body after the error and being charged.
Removed non_bots from GET /sentiment cohorts, where it repeated the main series point for point; read the non-bot population from series. This is a breaking change for integrations that read cohorts.non_bots.
Ranked people and topics above crypto tokens worth under $50 million that share their name, so a plain word such as tariffs finds the topic first.
Matched people and topics on the other names they go by, in the API and Nebula’s search bar, so tariffs finds Tariff, america finds United States and cryptocurrency finds Crypto.
Named the way past every plan limit: a history window beyond the plan, the rate limit and exhausted credits now say what the next plan allows and carry an upgrade_url, and the credits error gives the renewal date.
Answered requests that run past 60 seconds with a JSON 503, refunded, instead of an HTML page or a gateway timeout, and made GET /related-subjects over long windows for the most discussed stocks return in seconds.
The Developer API now tells every caller where it stands, streams Agent answers as they are written, and rejects mistakes before they are charged.
Added RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset to every response, X-Credits-Charged and X-Credits-Remaining to every fixed-price response, and GET /account, a free read of the plan, limits and credits left this cycle.
Streamed Agent chat completions as the answer is written, instead of in one piece at the end, with progress comments while tools run.
Rejected unknown parameters, unsupported filter values and malformed timestamps with a 400 that names the problem, before any credit is charged; integrations that sent parameters the API used to ignore now need to drop them.
Ranked GET /search the same way as Nebula’s search bar, so Bitcoin, SpaceX and people are no longer outranked by tokens trading on their names.
Added asset_id and asset_class, or subject_id, to every entity on insights and calendar events, so each can be queried directly.
Returned 404 for unknown paths and 405 for the wrong method instead of a plan error, accepted subject IDs in any letter case, and let GET /social-momentum return only the clusters you name.
The Developer API, MCP and CLI can now rank and screen what the market is talking about, by share of the conversation, how fast that share is growing, sentiment and price.
Added GET /top-trending and GET /top-momentum, which rank stocks, commodities, indices, crypto, or people and topics by their share of the last 24 hours of conversation and by the day-over-day change in that share.
Added GET /screener, which ranks by share, share change, posts, net bullish posts or price change within ranges on the others, and rejects unknown or inverted filters.
Let every ranking filter by author cluster, such as Smart Money or KOLs, measuring share against everything the selected clusters posted.
Renamed the home page treemaps to Top Trending and Top Momentum, and made their cluster views rank by the same mindshare definition as the unfiltered view.
Regrouped the API reference into Discovery, Lookup, Social & Sentiment, Market Sentiment and Market Data, and renamed the calls endpoints Long/Short Calls and Calls Outcome Distribution under Market Sentiment.
Replaced GET /clusters/{cluster_id}/assets with GET /top-trending?clusters=…, which also combines clusters and reports the day-over-day change; this is a breaking change for existing integrations.
Free and Pro now reach the same complete API, so plans differ only by history, rate limit and included credits.
Opened every endpoint and parameter to Free, including calls, insights, calendar events, conviction, price levels and cluster-filtered sentiment and mindshare.
Extended Free history from 7 to 90 days, and gave Pro full history.
Added 200,000 monthly credits to Pro, renewing every billing month at 25% below prepaid pricing, or 40% on yearly billing.
Returned a clear 403 before any credit is charged when a request reaches past the plan’s history window, and documented the limit on each affected parameter in the REST, MCP and CLI references.
The Developer API, MCP and CLI now search every stock, commodity, index, crypto asset, person and topic from one endpoint, treat commodities and indices as asset classes of their own, and explain every metric in plain English.
Added GET /search, which ranks assets, people and topics together and returns the IDs needed to query each one, replacing /resolve.
Returned GLD as asset_class=commodity and SPX as asset_class=index everywhere assets appear, and accepted those classes on every asset endpoint.
Rewrote every endpoint reference in plain English, explained the author clusters behind sentiment, mindshare and calls, and moved parameter rules into their own list.
Corrected single-asset sentiment, which counted part of the Smart Money cluster twice, and mindshare, which counted bot posts for the asset but not for the market, and matched cluster call counts to each cluster as it stood on the day of each call.
Rejected unknown cluster IDs with a clear error, moved importance scores to a 1–3 scale, and returned the same crypto asset IDs from cluster views as from asset search.
Retired the trading-volume and volatility endpoints and the volume series in price history, and replaced /resolve with /search; these are breaking changes for existing integrations.
A new pane under the price chart shows how much more bullish one group of authors is than another, opening on Smart Money against Dumb Money.
Added a Sentiment Bias pane that plots the gap between two author groups, switchable to any pair of clusters on either side.
Normalised each side by its own posting volume, so the pane reads a difference in conviction rather than a difference in how loud a group happens to be on that asset.
Held the series back where the evidence is too thin to support a reading, rather than drawing a single post as an extreme.
Added crosshair labels that name which side is more bullish, and controls to reorder the panes under the chart.
The Fear & Greed Index has been recalibrated against the two public crypto indices it gets read alongside, correcting a bias that made it read fearful during rallies.
Corrected a fault that made the index read too low in rising markets: on 1 September it now reads 68 against public readings of 69 and 72, where the previous version read 45.
Fixed the sentiment inputs, which had been treating days with no social data as though sentiment were exactly neutral, distorting the index on any asset with a shorter posting history.
Merged the two price-position inputs into one measure of where price sits in its 90-day range, so a new high no longer pegs that input and hides further gains.
Cut the average gap to the public reference from 8.8 to 3.2 points, and from 20.3 to 2.3 across greed conditions.
Published the updated component names and weights through the metadata on the public /fear-greed-index endpoint.
Crypto price charts now draw real candles — open, high, low and close per bar — the way the equity charts already do, with the current bar updating every minute.
Added true OHLC candlesticks to crypto price charts across the hourly, four-hour, twelve-hour and daily intervals.
Covered the assets people actually chart, from Bitcoin and Ethereum through to the mid-caps, with a year of history behind every one.
Kept the newest candle live: the bar currently forming updates every minute rather than waiting for the hour to close.
Left assets without an exchange listing on the existing line chart, so nothing lost detail it previously had.
The Developer API, MCP, and CLI can now show which non-asset subjects are being discussed alongside a stock or tracked topic.
Added related-subject discovery for equities and non-asset subjects, including shared posts, distinct authors, momentum, and the anchor's sentiment within each conversation.
Generated the new operation across REST, MCP, CLI, OpenAPI, and endpoint documentation from the same typed public contract.
Standardized subject discovery on Nebula's canonical names, aliases, merge resolution, and ranking while keeping raw registry classifications internal.
Simplified asset and subject results to stable identity fields and normal pagination, removing presentation-only image metadata and misleading total counts.
The Nebuliser now carries ninety days of history: pick a day from the strip along the bottom and the whole page — network, table and activity — redraws as it stood on that date.
Added a timeline under the network showing the subject’s daily post volume for the last ninety days, so a day worth reading is visible before you pick it.
Clicked any day to read the state of the network on that date, with the graph, the related-subjects table and both activity strips all moving to it together.
Kept the page live by default and one click away, so returning to the present is never more than pressing Back to live.
Carried the chosen day in the URL, so a network as it stood on a given date can be shared like any other view.
Labelled every mark with what it actually is — person, company, organisation, place, event — instead of where its record came from.
Every mark on the network is now drawn at the size of its conversation and placed by where that conversation is heading, so the subjects that matter are obvious before you read a single label.
Sized every subject by how many posts name it, so the biggest names on a network stand out from the tail instead of sharing its size.
Placed each subject by momentum rather than volume alone: surging subjects sit at the centre and fading ones move out to the rim, at any size.
Renamed the rings Surging, Rising, Steady and Fading to say what they now mean.
Named the marks in place — hover any subject to see what it is, and select one to name it and everything it is linked to at once.
Sent the expand control on a subject page's Topic Graph straight to the Nebuliser, where the same network is drawn at full size with its table and intelligence.
Carried the selected subject in the URL alongside the anchor, so a shared link opens on the exact pair you were reading.
The Nebuliser draws a stock, token or topic as the network of everything discussed alongside it, and lets you walk that network one subject at a time.
Added Nebuliser to the main navigation, drawing any subject's neighbourhood as a full-page network with its own search.
Ranked every neighbour by how much of the subject's conversation it holds, named the rings Core through Fringe, and linked the subjects that turn up in the same posts.
Selected a subject on the graph to light its links, scroll the related-topics table to its row, and read its share, momentum and sentiment.
Filtered the intelligence beneath the selection to the stories naming both subjects, so every link on the graph has its evidence beside it.
Nebulised any subject in one click to re-centre the whole view on it, with the anchor carried in the URL so a view can be shared.
Raised how many subjects a single view can hold, so the tail of the ranking is drawn rather than cut at the head.
A stock, token or topic page now opens with an AI overview of what is happening with it right now and why, written from the live conversation, price action and intelligence.
Added an AI Overview panel beside the charts on every asset and topic page, replacing the market-wide macro intelligence list that was identical on all of them.
Explained the move rather than restating it: the catalyst, the actors and the figures behind a change in price or attention.
Refreshed the top stocks, tokens and AI names on a schedule, and rewrote any subject immediately when a new insight names it, its attention spikes, or its price moves sharply.
Generated an overview on demand for any other page on first visit, so a quiet topic is covered the moment someone opens it.
Translated overviews, and the home page market digest, into every language Nebula supports.
The market digest now leads the home page and is written separately for equities, crypto and AI, so the day's stories match the market you are browsing.
Moved the three digest cards to the top of the home page, above the treemaps.
Generated a separate digest for Equities, Crypto and AI, each researched and written from its own market.
Switched the digest with the asset-class control, so choosing Equities loads equity stories rather than relabelled crypto ones.
Anchored each market on its own backdrop — rates, the Fed and earnings for equities, the AI buildout for AI, Bitcoin for crypto.
Extended the suggested questions on the chat bar to every market, where they were previously crypto-only.
Every stock, token and topic page now shows which other subjects turn up in the same posts, and what that does to the mood about the one you are reading.
Added a Discussed With table to every subject page, ranking the people, companies, stocks, tokens and topics that appear in the same posts.
Showed how many posts each pairing shares and what share of the conversation that is, counted once per post however many ways a subject is named.
Reported sentiment for the subject you are reading, measured only over the posts it shares with each of the others, against its own average for the period so you can see which company moves the mood.
Split every row bullish, neutral and bearish on the same cut used everywhere else in Nebula, so the numbers reconcile with the sentiment chart above them.
Made every row a link to that subject, and the window selectable from a day to a quarter.
People, companies, places, events and legislation now get the same sentiment dashboard as stocks and tokens, so you can follow a narrative that has no ticker.
Added a page for every tracked subject that is not a stock or a token, covering people, companies, organisations, places, concepts, technologies, events and legislation.
Carried over the units that do not need a price: sentiment, mindshare, post volume, emotions, time horizon, social momentum by cluster and intelligence.
Made the address forgiving — type a name and Nebula resolves it to the subject you meant, then moves you to its canonical page.
Unified the aggregates behind these pages so a person, an organisation and a topic are read the same way, and folding in more coverage later takes one change instead of three.
Every insight now shows the volume and mood of the conversation it came from, measured across the day around the story.
Added an activity chart to each insight, built from posts that discuss the story’s subjects together rather than separately.
Coloured each hour by its net sentiment, so a story’s reaction reads bullish, bearish, or flat at a glance.
Added a chatter mix showing which author clusters carried the story, and an impact readout scoring price, short-term, long-term, macro, and tail-risk exposure.
Added total mentions, distinct authors, and cluster reach beneath every insight.
Measured the window around the story itself, widening to the twelve hours either side once the reaction has landed.
Replaced the subject chips on insight cards, which repeated names already in the headline.
Live