category.sentiment_analytics.label

category.sentiment_analytics.promise

Starting path

Eight vendors sell social signal as a trading input. One publishes a test you can check.

Sentiment Analytics evaluated across published predictive evidence, published price and free-tier substance, documented bot filtering, and source and asset coverage.

category.sentiment_analytics.promise Reviewed on published predictive evidence, published price and free-tier substance, documented bot filtering, and source and asset coverage.

Start with the smart category assistant or jump straight into quick questions. Both paths narrow to the same recommendation standard.

How we review

Recommendation first, comparison only if needed.

Reviewed and rechecked on the current cycle.

Commercial relationships disclosed before clickout.

Audit

Infrastructure

Methodology

2026.04.0

published 2026-04-27

Providers tracked

710+

across 80 categories

Last verified

2026-08-07

editorial freshness

Editorial board

4

audit-log live

Decision guide

What matters most before choosing in this category

Weighted on published predictive evidence (30), published price and free-tier substance (25), documented bot filtering (25), and source and asset coverage (20). Across eight vendors, exactly one publishes a predictive test a buyer can check - and it is about US equities, not crypto. LunarCrush's Earnings Social Atlas covers 2,898 earnings prints with an event-level CSV downloadable without registration; recomputing it returns a decile spread of 1.186x, against the vendor's own companion note saying "The decile spread is modest by design" and "If you find a 5x effect, you have a bug, not alpha." The web page headlines a 1.2 to 1.4x range using an internal liquidity cut not included in the public file. That gap - between the checkable number and the advertised one - exists at the most rigorous vendor in the category.

Key question

Has this vendor published any test of whether its signal predicts anything?

Key question

What fraction of ingested posts does the bot filter actually remove?

Key question

Can you read the price, and does the free tier survive contact with real use?

Current editor lead

LunarCrush

Social feed aggregation across six platforms sold as Galaxy Score and AltRank composites plus an API and MCP server.

Data checked Aug 2026

Social feed aggregation across six platforms sold as Galaxy Score and AltRank composites plus an API and MCP server. Strongest on published predictive evidence (8/10): The only vendor in this pool publishing a test a buyer can reproduce, and I reproduced it. The companion guide states the dataset is "One row per US earnings print, January through June 2026: **2,898 events** across **13 sectors**" and gives the methodology: "z-scores are computed **against each name's own baseline**:… Weakest on published price and free-tier substance (5/10): Corrected during verification. Three different figures for the Individual plan ship inside the same JavaScript bundle: the plan card carries price:"$72" with no billing period attached, an unserved page-metadata field says "Individual at $90/mo", and a support FAQ answers "Plans range from free Discover tier to… Published price: Fetched 2026-08-07. Corrected during verification: https://lunarcrush.com/pricing serves a shell whose only rendered text is "Select a Plan" and "Choose a plan that fits your needs and unlock the right level of social intelligence." - no price is…

Best for

Published predictive evidence — 8/10Source and asset coverage — 8/10

Main tradeoff

The one reproducible test in this pool is on 2,898 US equity earnings prints in 2026 H1, not on crypto, and the effect that actually reproduces from the public file is about 1.19x top-versus-bottom decile. Corrected during verification: the vendor's own companion file discloses that the louder figures on the Atlas web page - "~1.2–1.4×", "+1.79σ print-day average", "1,665 liquid prints", "29.7% of liquid prints build more than 1σ of attention before the number is public" - are computed on an internal panel using a dollar-volume liquidity cut "which is not included in this public file", and warns to "Expect directionally identical but numerically different results". So the headline a buyer sees is not the number he can check; the number he can check is smaller. The page also routes the tradeable version - hourly feed, point-in-time constituent membership, 5+ years of history - to a two-week institutional trial arranged with the CEO by email. A crypto buyer is extrapolating an equities result across an asset class, a frequency and a price tier none of which were tested in public.

Verify before signup

Fetched 2026-08-07. Corrected during verification: https://lunarcrush.com/pricing serves a shell whose only rendered text is "Select a Plan" and "Choose a plan that fits your needs and unlock the right level of social intelligence." - no price is server-rendered, and the page's served <meta name="description"> is that same sentence. Every figure below was read from the JavaScript bundle the page loads, which is this pricingSourceUrl. It contains pricingPlans:[{name:"Hobby",description:"Track prices, market caps, and basic market data for thousands of crypto assets and stocks. No social analytics or API access.",price:"",priceLabel:"FREE"...},{name:"Individual",...price:"$72"...},{name:"Builder",...price:"$240"...},{name:"Scale",...price:"$720"...},{name:"Enterprise",...priceLabel:"Contact Sales"...}]. None of those price fields carries a billing period. The same bundle holds an unserved metadata field meta_description:"Compare LunarCrush plans. Individual at $90/mo, Builder at $300/mo, Scale at $900/mo. All plans include MCP access for Claude, ChatGPT, Gemini and more." and a support FAQ answer "Plans range from free Discover tier to Individual ($24/month), Builder ($240/month), and Enterprise (custom pricing)." Corrected during verification: the draft called the $90 figure the pricing page's meta description; it is not served as such. Note also that $72, $240 and $720 are each exactly 0.8x the $90, $300 and $900 metadata figures, consistent with an undisclosed 20% annual discount rather than a pure contradiction.

Weighted criteria

Published predictive evidence30%

Published price and free-tier substance25%

Documented bot filtering25%

Source and asset coverage20%

Leading options

Shared shortlist for this category

These providers are pulled from the same category comparison catalog used in validation, so the category page, comparison page, and provider reviews stay aligned.

LunarCrush

Social feed aggregation across six platforms sold as Galaxy Score and AltRank composites plus an API and MCP server.

Editor lead

Best for: Published predictive evidence — 8/10

Santiment (Sanbase / SanAPI)

Social, on-chain and development metrics over seven enumerated text sources, sold as a dashboard plus API.

Social, on-chain and development metrics over seven enumerated text sources, sold as a dashboard plus API. Strongest on source and asset coverage (10/10): The most completely enumerated coverage in the pool - every source named with its scale, its latency and the date its history begins: Telegram "over 400 crypto-related Telegram chats" at "1-2s delay max" from 2016-03-29; Twitter "over 4000 crypto-related and NFT-related Twitter accounts" from 2018-02-13; Reddit "over… Weakest on documented bot filtering (5/10): Documents a real filter, but it is a classifier-confidence threshold rather than account-level bot detection. The page says that besides ordinary messages "some might be just neutral or contain spam or other irrelevant information", that "These kinds of messages usually have a pretty vanished pair of sentiment… Published price: Fetched 2026-08-07 from the pricing page. Corrected during verification: the draft joined separate table cells with slashes into single quoted strings; the contiguous published text is quoted here instead. The Free plan is "Ideal for new traders and investors…

Best for: Source and asset coverage — 10/10

The Tie (Terminal and Data APIs)

Institutional terminal and quant API whose sentiment product is derived entirely from Twitter, sold only through sales.

Institutional terminal and quant API whose sentiment product is derived entirely from Twitter, sold only through sales. Strongest on documented bot filtering (6/10): Corrected during verification: the draft cited the marketing page for all three strings, but the two field definitions are not on it - I searched https://www.thetie.io/solutions/api and neither the field names nor their definitions appear there. The marketing assertion "Bot and spam detection to ensure clean signal"… Weakest on published price and free-tier substance (1/10): Lowest possible transparency: no number on any of the pages loaded on 2026-08-07, no free tier, no trial, no self-serve signup. Access to the data requires an API key that "is unique to data access and will not be displayed in your Tie Terminal admin page" and is issued by an Account Manager, per this page's… Published price: No price is published on any The Tie page loaded on 2026-08-07 (home, /solutions/terminal, /solutions/api, /legal/disclosure, /insights, docs.thetie.io and the API reference). The Terminal page's only commercial calls to action are "Schedule a demo" and…

Best for: Documented bot filtering — 6/10

Browse this network

Methodology

How this category is reviewed

Reviewed on published predictive evidence, published price and free-tier substance, documented bot filtering, and source and asset coverage.

Reviewed on: Published predictive evidence, Published price and free-tier substance, Documented bot filtering, Source and asset coverage.

This page is a maintained category surface, not a static marketing block. Review freshness, provider positioning, and recommendation logic should stay consistent with quiz and provider pages.

Frequently asked

Questions people ask before choosing sentiment

Does any sentiment vendor prove its signal predicts price?

One, partially, and not for crypto. This was tested by looking rather than assumed: The Tie's /research and /resources return HTTP 404 and its live insights index carries roughly a hundred items - funding briefs, acquisitions, event recaps, webinars - none reporting a correlation, hit rate or accuracy figure. Santiment makes no predictive claim at all. Kaito publishes a one-sentence definition of mindshare and no methodology page. alternative.me states two assumptions and validates neither. cookie.fun documents features only, and Augmento's current site contains no instance of "research", "correlation", "Spearman" or "study". That leaves LunarCrush's Earnings Social Atlas, which is real, checkable, and about US equity earnings prints. StockGeist is the near-miss: its evidence page promises "evidence that StockGeist has correctly forecast market fluctuations", and its AYTU case study describes social activity that "followed this announcement immediately" after a company press release - the signal trailing the news, offered as forecasting, with no sample, no counterfactual and no accuracy figure.

How much of the feed is bots?

No vendor here says. Five assert filtering - LunarCrush "removes automated posts, duplicate-network activity, and low-quality content", Kaito "extensively leverages LLM to exclude low-effort content including spams", The Tie offers "Bot and spam detection to ensure clean signal", Santiment applies a score threshold, and cookie.fun flags "signs of botted likes or views" - and not one states what fraction of ingested posts the filter removes, what it catches, or what it misses. Searching LunarCrush's entire shipped JavaScript bundle for a removal percentage, a removed count or a filtered-out figure returned none. Two vendors do ship a proxy you can compute yourself: LunarCrush exposes spam as a queryable time-series metric to "gauge manipulation risk", and The Tie exposes tweet_unique_sources_percent. A proxy you compute is not a disclosed removal rate, but it beats an assertion.

Is there a conflict of interest to watch for?

At one vendor, yes, and it is disclosed rather than hidden. The Tie scores sentiment on tokens while The Tie Capital LLC - per its own disclosure dated 28 May 2026 - "is a broker-dealer registered with the U.S. Securities and Exchange Commission" and a "member of the Financial Industry Regulatory Authority", providing "private capital raises, mergers and acquisitions advisory" to clients in the digital asset industry. The disclosure covers banking risks and does not address information barriers between the bank and the sentiment feed. That is not an accusation; it is a question a buyer should ask before treating the feed as neutral.

REVIEWEDApr 2026METHOD4 criteriaCATEGORYsentiment_analytics

Not financial advice · For informational purposes only · Always do your own research

Octopus · The AI CFO that pays for itself

Audit your entire crypto stack — free.

Score concentration, fees, security, and tax complexity across exchanges, wallets, staking, and DeFi.

Not financial advice · For informational purposes only · Always do your own research