NFT Analytics evaluated across what the vendor removes before it prints the number, what the printed number actually counts, a mark for the token you hold that comes with a published error, and whether the number is revised after you have used it.
Ranked here: NFT analytics vendors that publish a number for an NFT, such as a collection floor, market cap or per-token appraisal, for a reader who must defend it. Not ranked here: marketplaces where an NFT is bought or sold, general onchain research platforms, and fast rarity ranks for sniping listings, since no criterion here measures speed.
#1 of 6 · published ranking
NFTGo
78ChainChoice Score
4199
Why it leads
NFTGo is 3 points ahead of bitsCrunch (UnleashNFTs), mainly because it scores higher on its revision policy.
Show the arithmetic
Ahead of bitsCrunch (UnleashNFTs) on revision policy (+5.0 points)
Cost
Not priced· No comparable price is published
ReviewedNext review by Published Methodology v2026.11.1
How this page is reviewed
Reviewed means every product ranked on this page was read against its sources on or after that date; every ranked category is re-read at least every 45 days. Models research, extract and parse. A deterministic engine ranks. Every quote passes an adversarial refute pass.
No provider can pay for a position in this table. The code that computes this order cannot read which links carry a referral code, and every build proves it. Every score below rebuilds from the published criteria.Ranking v2026.11.1 · 4 criteria · 6 products · same inputs, same order
How we review · 6 ranked of 6 assessed · 4 criteria · 0 sponsor weight
How we review
6ranked
6assessed
4criteria
0sponsor weight
Recommendation first, comparison only if needed.
Reviewed and rechecked on the current cycle.
Commercial relationships disclosed before clickout.
Personal comparison
NFT Analytics ranked comparison
No comparable price is published; ranking still uses verified product evidence.
RankProviderBest fitScore
1
NFTGoAn NFT data API in two published versio...Top ranked
What the vendor removes before it prints the number — 6/10Trade-off: FIRST, the wash-filter rule is published: https://docs.nftgo.io/docs/wash-trade-filte...
78score
Check availability for your country
NFTGo · docs.nftgo.io. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
2
BIbitsCrunch (UnleashNFTs)A credit-metered REST API (66 documente...
A mark for the token you hold, with a published error — 7/10Trade-off: The two headline claims in the pool record did not survive the docs. bitsCrunch does...
75score
Check availability for your country
bitsCrunch (UnleashNFTs) · unleashnfts.com. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
What the vendor removes before it prints the number — 5/10Trade-off: Not established in this review: 1) ACCESS ROUTE, DISCLOSED SO A VERIFIER CAN REPRODUC...
66score
Check availability for your country
CoinGecko NFT · coingecko.com. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
4
NFTScanA multi-chain NFT indexer and explorer...
What the printed number actually counts — 6/10Trade-off: NFTScan's docs SPA (docs.nftscan.com/guides/...) renders its prose client-side from r...
61score
Check availability for your country
NFTScan · nftscan.com. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
Same score, not joint · ordered by weighted total (34.90 against 34.80)
What the vendor removes before it prints the number — 6/10Trade-off: Both API hosts CryptoSlam publishes are unreachable, and the one that works is undocu...
61score
Check availability for your country
CryptoSlam · cryptoslam.io. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
6
NFT Price FloorA free, ad-supported web ranking of NFT...
Whether the number is revised after you have used it — 5/10Trade-off: Two distribution channels publish different commercial terms, and they cannot be reco...
56score
Check availability for your country
NFT Price Floor · nftpricefloor.com. Provider links appear once we know which country you are browsing from: ChainChoice links to providers only where it serves in full.
Commission: This provider pays ChainChoice nothing for this link. Disclosure
Crypto products are high risk. Check the provider is authorised where you live before you deposit.
Ranked on 4 published criteria weighted 30/26/24/20, which set goal alignment — 30 of the 86 points. The rest: regional access 20 (constant here), evidence depth 18 (constant here), ease of use 10 (constant here), institutional trust 8 (constant here). Profile match is shown in breakdowns but carries no weight.
How you can lose money here, and what the ranking scores
Information failure: the number is wrong or bought, or your data leaks.
Scored by What the vendor removes before it prints the number · 30% (NFTGo 6/10, bitsCrunch (UnleashNFTs) 6/10, CoinGecko NFT 5/10)
Scored by What the printed number actually counts · 26% (NFTGo 5/10, bitsCrunch (UnleashNFTs) 5/10, CoinGecko NFT 5/10)
Scored by A mark for the token you hold, with a published error · 24% (NFTGo 6/10, bitsCrunch (UnleashNFTs) 7/10, CoinGecko NFT 2/10)
Scored by Whether the number is revised after you have used it · 20% (NFTGo 5/10, bitsCrunch (UnleashNFTs) 1/10, CoinGecko NFT 2/10)
What matters most before choosing in this category
Weighted on what the vendor removes before it prints the number, what the printed number actually counts, a mark for the token you hold, with a published error, and whether the number is revised after you have used it, in that order. THE FILTER EVERY VENDOR ADVERTISES IS SWITCHED OFF EXACTLY WHERE THE NUMBER LEAVES THE BUILDING. Of the six vendors ranked here, exactly one removes inorganic trades from the printed figure by default (CryptoSlam's aria-label tooltip, served on https://www.cryptoslam.io/, /nfts, /collections and /blockchains: "Wash sales are NOT included in Sales (USD) aggregation."). NFTGo's filter is default-ON in the web product for an enumerated 55-entry metric table (https://docs.nftgo.io/docs/wash-trade-filter-1: "the wash trade filter has been automatically turned on for the following key metrics") and default-OFF in the API a buyer actually integrates — the published OpenAPI parameter block reads {"name":"exclude_wash_trading","schema":{"type":"boolean","default":false}}, and the vendor's own tutorial "How to filter out wash trades?" ships its copy-paste sample as "const excludeWashTrading= false;".
Who is going to challenge this number — an auditor, a lender’s risk desk, or a tax authority?
Are you marking a whole collection, or one specific token that is not the cheapest in it?
If the vendor revises its filter next quarter, can you still retrieve the figure you actually used?
Leads on disclosure
The #1 overall and the scoring weights
NFTGoData checked Sep 2026
An NFT data API in two published versions — a stable V1.1 on Ethereum only, and a newer V2.2 "Multi Chain API" covering eight mainnets — plus GoPricing, a per-token appraisal that returns an estimated price for a specific tokenId decomposed into a collection floor base, an intercept and machine-learned per-trait premiums, and returns the trait weights and intercept themselves so the buyer can re-derive the number. GoPricing is documented only under V1.1 and is Ethereum-only. Wash-trade filtering is default-on in the web product for an enumerated table of metrics but default-OFF in the V1.1 API (exclude_wash_trading defaults to false) and has no exclude parameter at all on the V2.2 transaction-history endpoint, which instead tags each last sale with an is_wash_trade boolean. No accuracy figure, model version or as-of query is published anywhere.
An NFT data API in two published versions — a stable V1.1 on Ethereum only, and a newer V2.2 "Multi Chain API" covering eight mainnets — plus GoPricing, a per-token appraisal that returns an estimated price for a specific tokenId decomposed into a collection floor base, an intercept and machine-learned per-trait premiums, and returns the trait weights and intercept themselves so the buyer can re-derive the number. GoPricing is documented only under V1.1 and is Ethereum-only. Wash-trade filtering is default-on in the web product for an enumerated table of metrics but default-OFF in the V1.1 API (exclude_wash_trading defaults to false) and has no exclude parameter at all on the V2.2 transaction-history endpoint, which instead tags each last sale with an is_wash_trade boolean. No accuracy figure, model version or as-of query is published anywhere. Strongest on what the vendor removes before it prints the number (6/10): NFTGo's docs say its wash-trade filter is on by default for key web metrics, removing trades found by 'transaction cycles and transaction closed loops', but state no window or threshold, and its V1.1 API sets exclude_wash_trading to false by default. Weakest on what the printed number actually counts (5/10): NFTGo defines the floor as 'the lowest-priced item for sale' and publishes a three-branch market-cap formula, but names its venues only as 'notable marketplaces' and says nothing about bundles, sweeps or private sales. Published price: No price is published on any page loaded on 28 September 2026. The Compute Units page at https://docs.nftgo.io/reference/compute-units (HTTP 200) states: "To ensure more reasonable pricing for our services, we utilize "compute units" (CU) to describe the cost for…
Best forWhat the vendor removes before it prints the number — 6/10
Main tradeoffFIRST, the wash-filter rule is published: https://docs.nftgo.io/docs/wash-trade-filter-1 carries a 'Wash Trade Auto Filter' table enumerating the metrics filtered automatically in the web product, and both metric glossaries label each affected figure 'Suspected wash trades have been filtered out'. The rule itself is still unreproducible ('transaction cycles and transaction closed loops'), and the V1.1 API default is false in the published OpenAPI: default-on in the UI, default-off in the API, both series recoverable. SECOND, on accuracy: the words 'accuracy, stability and Interpretability' are a link description in the 'Learn More' list on https://docs.nftgo.io/reference/gopricing, not the substance of the evaluation docs; https://docs.nftgo.io/reference/evaluation-method names MAPE as the primary metric and publishes its formula, omitting only every number. THIRD, it is not a flat contradiction that algorithm-principle says 'Parameters are calculated by linear regression according to transaction history' while gopricing says 'Normally, gradient-boosting models perform well in small data scenarios of the NFT market.' Read in context the gopricing line is a general remark — it is preceded by 'This floor price-based framework is suitable for any regression model' — and the Stanford Blockchain Review paper NFTGo links from that same page discusses GBDT at length specifically to reject it for poor interpretability in favour of the linear Premium Model. So this is an ambiguity about which estimator is actually fitted, not two flatly opposed claims; it remains a disclosure defect because no page states the shipped estimator unambiguously. Finally, the formula 'Estimate Price = Floor Price + Intercept + Sum of (Trait Weights x Floor Price)' is rendered on the algorithm page as an IMAGE (files.readme.io/710898f-image.png), not as text (as the page's markdown source shows). An automated summariser returned it to an earlier pass as a verbatim quote. It is not quoted as text anywhere in this record and no downstream copy should either.
Verify before signupNo price is published on any page loaded on 28 September 2026. The Compute Units page at https://docs.nftgo.io/reference/compute-units (HTTP 200) states: "To ensure more reasonable pricing for our services, we utilize "compute units" (CU) to describe the cost for each API on NFTGo, as the cost can vary significantly across different APIs. Your charges will be based on the volume of CUs consumed." It then publishes per-endpoint CU costs under the columns "API Endpoint | CU(s) | CU(s) Additional" (for example, Get NFTs by collection costs 800 CU and Get rarity by NFT 100 CU) but no currency amount per CU and no plan tiers. The docs name the plan and billing surfaces elsewhere: https://docs.nftgo.io/v2.2/reference/introduction.md says "You can view [Pricing](https://developer.nftgo.io/pricing) and choose the package plan that best suits your needs" and "Currently, we have a rate limit of 5 requests per second (5RPS) for a free trial." On 28 September 2026 https://developer.nftgo.io/pricing and https://developer.nftgo.io/ both returned HTTP 404 with the body {"error_msg":"404 Route Not Found"}, exactly as on 14 August 2026, and https://nftgo.io/pricing returned HTTP 403 behind a Cloudflare challenge (a rendering fetch was also refused with 403). No CU-to-currency rate is readable anywhere, so cost cannot be compared before a sales conversation.
Weighted criteria
What the vendor removes before it prints the number30%
What the printed number actually counts26%
A mark for the token you hold, with a published error24%
Whether the number is revised after you have used it20%
Methodology
How this category is reviewed
Reviewed on what the vendor strips out before it prints the figure, what that figure actually counts, whether it prices the specific token you hold and admits its error, and whether it restates history after you have used it. Weights are the engine’s: 30 / 26 / 24 / 20.
The order on this page is the published ranking for this category. Every criterion, weight and source behind it is on the methodology page.
Frequently asked
Questions people ask before choosing NFT data
Why can two vendors quote the same collection and disagree?
Because a floor price is not observed, it is constructed, and each vendor constructs it differently. NFT Price Floor advertises that it aggregates “all marketplaces”, but across all 1,799 records served by its own backend, 93.2% of priced collections are marked off a single venue. Four of the six vendors here also print the words “market cap” and compute incompatible figures from them — floor price times total supply at one, the average of the last 50 trades times item count at another — and none states which formula it used at the point it hands you the number.
Does the wash-trading filter apply to the number I actually integrate?
Usually not. Of the six vendors ranked here, exactly one removes inorganic trades from the printed aggregate by default. NFTGo turns its filter on for the website and ships the API parameter exclude_wash_trading with a default of false, so the figure on the page and the figure in your NAV pack are different numbers for the same collection — and the unfiltered one is the one that reaches your report. Ask which surface the filter is on for, not whether the vendor has one.
Is a collection floor good enough to mark the token I hold?
Only if your token is the cheapest one in the set, because that is what a floor is. Four of the six vendors publish no per-token output at all, and two say so outright — CoinGecko: “At the moment, our API only supports the market data for the entire NFT collection.” Of the two that do price a specific token, only bitsCrunch ships an admitted uncertainty alongside it, returning lower and upper bounds with every estimate. NFTGo publishes the construction of its model but no accuracy figure, while its marketing carries an “over 90%” its own documentation never defines.
Can I retrieve what a vendor published last quarter?
No vendor in this pool offers an as-of query that returns the figure as it stood when you used it. NFTScan states the consequence plainly: “NFTScan doesn’t store historical snapshots of the floor price of a collection, while it only updates the current floor price of the collection.” That is the difference between a historical series recomputed today and a record of what you were told — and it is why a filing defended two years later can stop being reproducible at exactly the moment it is examined.
Reviewed28 September 2026Method4 criteriaCategorynft_analytics
Not financial advice · For informational purposes only · Always do your own research
