Skip to content

Methodology

How we track sentiment

Every source we track, how we label sentiment, and how the record reaches your watchlists and your AI.

Active sources
2,800+
Channels
17
Records
2.7 million
Record starts
2011

Measured 27 Sep 2026

Sources

Complete source directory

2,800+ sources across 17 channels. Each count is the number of sources that published at least once in the 12 months to 26 Sep 2026.

X
1,125
accounts
Executives, analysts, developers, journalists, and organizations.
LinkedIn
421
accounts
Company pages and industry voices.
Mainstream media
240
outlets
Global news outlets and broadcast networks.
Podcasts and video
199
shows and channels
Transcribed podcast episodes and video shows.
Code repositories
189
repositories
Bitcoin, Ethereum, and other protocol and wallet repositories.
Financial media
139
outlets
Markets, business, and finance publications.
Company newsrooms
120
companies
Blogs and newsrooms published by the companies themselves.
SEC filings
109
companies
EDGAR filings from tracked companies.
Crypto media
55
outlets
Crypto-native publications covering digital assets.
Research
52
sources
Industry research, on-chain analytics, and papers.
Tech media
52
outlets
Technology publications covering digital assets and AI.
Regulators and policy
51
sources
Regulators, central banks, and policy bodies.
Conferences
37
event series
Talks and panels from industry conferences.
Newsletters
15
newsletters
Independent industry newsletters.
Data centers and energy
13
outlets
Mining, data center, and energy coverage.
Press releases
7
wire services
Business Wire, PR Newswire, GlobeNewswire, Chainwire, and others.
Communities
6
platforms
Reddit, Substack, Medium, Hacker News, Stacker News, and InvestorsHub.

Collection

Everything gets captured

Every mention across 2,800+ sources. Dedicated coverage, social posts, passing references. This powers your watchlist alerts and trend clustering in Spaces.

Curation

Broad and deep

RSS feeds, social accounts, regulatory filings, press wires, and research, all flowing through automated pipelines into a unified data warehouse. If it's published, we capture it.

Enrichment

AI-enriched for deep analysis

Every article is sentiment-scored, entity-extracted, and topic-classified using large language models. Trends are identified by grouping related narratives with signal strength and confidence scoring.

Delivery

From monitoring to deliverables

Create watchlists for what matters. Organize in Spaces by project. Generate deliverables with Recipes. All backed by verified coverage.

Labeling

How we label sentiment

We categorize by tone, context, and meaning, and every label follows the same written criteria.

Positive

Content that portrays Bitcoin favorably or highlights its benefits, adoption, or potential.

Signals: Endorsements, adoption announcements, bullish analysis, success stories, institutional support

Negative

Content that portrays Bitcoin unfavorably, highlights risks, or expresses skepticism.

Signals: Criticism, regulatory warnings, security concerns, environmental arguments, scam associations

Balanced

Content that presents multiple perspectives neutrally without advocating for a particular view.

Signals: Factual reporting, educational content, both-sides analysis, price updates without commentary

What Perception tracks

Coverage across the digital asset industry

  • Bitcoin

    The base layer: institutional commentary, developer activity and regulatory narratives.

  • Stablecoins

    Adoption trends, issuer moves and regional policy shifts.

  • Tokenized assets

    Real-world asset tokenization, market developments and regulatory frameworks.

  • Macro catalysts

    Rate moves, FX flows, capital controls and other drivers of digital asset narratives.

Query the scored record through MCP