Stock sentiment from unstructured news content with LLMs.

Contact us

News are noisy. Get signal for your quantitative strategies.

01
cONTENT
fingerprint

Full content analyzed.

Full news article content (not just the article title) analyzed for relevancy to the stock and sentiment.
02
TICKERS
fingerprint

Tickers mapped.

OpenFIGI used for ticker mapping. Additional proprietary mapping methods of stocks to news article content.
03
SENTIMENT
fingerprint

State-of-the-art sentiment with AI.

Granular intraday sentiment on 11-point scale by a custom Large Language Model (LLM).
04
· MULTI-LINGUAL
fingerprint

41 languages.

News content from 41 languages including Chinese, Japanese, Arabic, Persian, German and all major languages.
05
Global coverage
fingerprint

1,000,000+
articles / month.

Analyzing business news from publicly available trusted global sources with links.
06
DELIVERY
fingerprint

Delivered via API. JSON format.

Data pushed in real-time to you via our API.
07
BACKTESTING
fingerprint

Backtesting data.

Datasets available for past 4-5 years covering 1,000s of stocks and various markets (e.g. US, Brazil, Japan). Custom markets available.

Customers.

"In the investing world, having an information edge is everything. With Reclever.ai we can parse through unstructured data faster and find relevant signal that matters."

Benefits.

check

Unlock Hidden Alpha.      

Our API uses custom LLMs to provide granular, 11-point sentiment scores from unstructured news data. This helps your firm identify subtle market shifts and uncover unique alpha-generating opportunities.
check

Easy To Integrate.

Get backtestable datasets for the US and Brazil, all mapped with OpenFIGI and delivered in JSON via API. Easy integration and global coverage you need to refine your strategies.

Events we attend.

FAQ

check
Question: What kind of content does the Reclever API analyze?

Answer: Our API analyzes the full content of news articles, not just the headlines, to determine relevancy to a stock and its sentiment. We analyze over 1,000,000 articles per month from trusted global sources.
check
Question: How is sentiment measured?

Answer: We use a custom Large Language Model (LLM) to provide granular intraday sentiment on a detailed 11-point scale, offering a more nuanced view than simple positive/negative scores.
check
Question: What markets and languages are covered?

Answer: Our API analyzes news content from 41 languages, including Chinese, Japanese, and Arabic. We provide specific backtesting datasets for US and Brazil markets, with custom markets available.
check
Question: What is alternative data in the context of financial markets?

Answer: Alternative data refers to non-traditional data sources that can provide an information edge in financial analysis. For Reclever.ai, the primary source of alternative data is the full, unstructured text of global news articles. By analyzing this data, we uncover signals and sentiment shifts that are not apparent in standard financial statements or market price data.
check
Question: How does Reclever.ai measure news sentiment for financial analysis?

Answer: Reclever.ai measures news sentiment using a custom-built Generative AI model. Instead of a simple positive, negative, or neutral score, we provide a granular 11-point sentiment scale. Our model analyzes the full content of an article, not just the headline, to determine its relevance and true emotional tone concerning a specific financial asset. This provides a more nuanced and accurate signal for investment strategies.
check
Question: What is the role of Generative AI in Reclever's news sentiment analysis?

Answer: Generative AI, specifically our custom Large Language Model (LLM), is at the core of our service. It allows us to move beyond keyword-based analysis and truly understand the context, subtleties, and nuances of financial news in over 41 languages. This technology enables us to parse millions of unstructured articles and convert them into structured, actionable sentiment data that quantitative firms can use to identify unique, alpha-generating opportunities.
check
Question: Why is analyzing the full text of a news article better than just the headline?

Answer: Headlines can be misleading or lack crucial context. The true sentiment and impact of a news event are often detailed within the body of the article. Our Generative AI analyzes the full text to ensure the sentiment score is based on a comprehensive understanding of the information, leading to a more robust and reliable signal.
check
Question: How is the sentiment data made actionable for trading?

Answer: The sentiment data is delivered via an easy-to-integrate API in JSON format. All data is mapped to specific financial instruments using the OpenFIGI standard for ticker symbols. We also provide historical, back-testable datasets for markets like the US, Japan and Brazil, allowing firms to validate their data against their trading strategies.
check
Question: How is news sentiment data practically used for investing?

Answer: Investment managers use news sentiment data as a key input for their decision-making process. It can serve as an early warning system for negative events, a confirmation signal for a positive thesis, or a way to identify undervalued assets that have positive but yet-unrecognized sentiment momentum. For example, a rising sentiment score for a company's products, even before sales numbers are reported, can be a leading indicator of future growth.
check
Question: What is the specific application of news sentiment in quantitative trading?

Answer: In quantitative trading, news sentiment is used as a predictive feature in algorithmic models. Quants use historical sentiment data, like our backtesting datasets, to find statistical correlations between sentiment shifts and subsequent stock price movements. A model might be designed to automatically long stocks with a sustained high sentiment score while shorting those with a sudden, sharp decline in sentiment, all executed systematically at high speed.
check
Question: How can sentiment data generate 'alpha'?

Answer: Alpha represents investment returns that are not attributable to general market movement. Sentiment data can generate alpha by providing an 'information edge'. Because our AI analyzes the full text of news articles in near real-time, it can capture nuanced market perception shifts faster than human analysts and before that information is fully reflected in the stock price. This time advantage allows trading strategies to act on new information first, which is a classic source of alpha.
check
Question: What are common sources of alpha for fund managers?

Answer: Fund managers seek alpha from various sources. These include informational advantages (knowing something before the market does), analytical advantages (interpreting public information better or faster), and behavioral advantages (exploiting market irrationality). High-quality alternative data, like the granular news sentiment provided by Reclever.ai, directly contributes to analytical advantages.
check
Question: Why is a granular, 11-point sentiment scale important for generating alpha?

Answer: A simple positive/negative score is a blunt instrument. A granular 11-point scale allows quantitative models to detect subtle but significant changes. For instance, a shift from 'very positive' to just 'positive' can be a powerful leading indicator of decelerating momentum, even though the sentiment is still positive overall. This nuance is often where alpha is found.