Your guide to market sentiment.

See what the financial press really thinks. Sentiment Explorer reads analyst articles as they are published, extracts every buy, hold and sell call, and turns them into one weekly picture of the market's mood.

Explore the charts MCP How it works

Ask your AI agent

Bring the whole corpus into your chat.

Sentiment Explorer is also an MCP server. Connect it to Claude or ChatGPT and ask in plain words: your agent reads the sentiment and its trend, the buy, hold and sell split, how each source leans, and the article summaries, then checks them against the news. Free and read-only, no account needed.

Ask about one company.

/sentiment-explorer what's up with Nvidia?

  • Overall tilt is positive. 46 articles with 46 recommendations: 29 buy, 17 hold, 0 sell. The average sentiment is +4.2 on a −10 to +10 scale, and the quality-weighted figure is the same. That is above the corpus-wide average of +3.0.
  • The trend has dipped and partly recovered. Late July to August was strongly bullish, at +5 to +7 and nearly all buys. In the week of Sep 7 the average dropped to +2.9, with 9 of 13 calls at hold, around the mid-September chip sell-off on AI-slowdown fears. October recovered to about +4.5 to +5.5, and the latest week is +3.0.

Ask why the mood moved.

/sentiment-explorer what are the reasons of current stock sentiment?

Sentiment sagged through August and September, climbed to a peak in the week of Oct 12, then eased to +3.05 this week, with holds rising as a share of recommendations. Part of that Oct 12 peak is one dividend-screen article from Oct 14 that carried about 20 “buy” calls on its own. Treat the spike as somewhat inflated.

Set it up in three steps

  1. Add the connector In Claude open Settings → Connectors → Add custom connector. Name it “Sentiment Explorer”, paste the address below, leave the OAuth fields empty and select Add. https://mcp.sentiment-explorer.com
  2. Download the skill The skill teaches the agent how to research with the connector: which tools to use, how to weigh the sources and how to cite the articles. Download skill (.zip)
  3. Install the skill Turn on Settings → Capabilities → Code execution and file creation. Then open Customize → Skills, select + → Create skill → Upload a skill and choose the zip. Start a chat with /sentiment-explorer and your question.
  1. Add the connector In ChatGPT on the web turn on Settings → Security and login → Developer mode. Then open Plugins, select + → Create an MCP app, name it “Sentiment Explorer”, paste the address below, set Authentication to No authentication, confirm and select Create. https://mcp.sentiment-explorer.com
  2. Download the skill The skill teaches the agent how to research with the connector: which tools to use, how to weigh the sources and how to cite the articles. Download skill (.zip)
  3. Install the skill Open Plugins → Skills, select Create → Upload from your computer and choose the zip. In a new chat, turn on Sentiment Explorer from the + menu and ask about a stock, an ETF or a theme.

How it works

From thousands of articles to one line.

  1. ReadAnalyst articles from financial news sites, collected as they are published.
  2. ExtractA language model pulls out every recommendation: the instrument, the call, and a sentiment score from −10 to +10.
  3. AggregateScores, themes and mentions are rolled up by week, so a shift in mood shows up as a shift in the line.