How Crawlux scores a crypto site.
Open methodology is verifiable methodology. This page documents the 23 analyzers, the 6 check groups, the scoring weights, where the calibration data comes from, the versioning policy and how to challenge a score. The audit logic is opinionated where it needs to be and open where it can be.
Why the methodology is open
Closed audit methodologies are black boxes. Buyers cannot verify the logic. Findings feel arbitrary. Disagreements have no path to resolution. The Crawlux methodology is the opposite: every weight, every threshold, every calibration source is documented.
Closed methodology
What most audit tools do
- Proprietary scores with no published rules
- No calibration sources disclosed
- No version history or change log
- No path to challenge a score
- Black box for buyer evaluation
Open methodology
What Crawlux publishes
- Every analyzer documented with thresholds
- Calibration sources cited per dimension
- Quarterly version history tracked
- Email path to challenge findings
- Verifiable for buyers and auditors
The moat is execution, not the rules
Other auditors are welcome to use the Crawlux methodology with or without Crawlux. The competitive advantage is running 23 analyzers across 30+ AI prompts in 2 minutes, not the rules themselves.
The 6 check groups
Every analyzer belongs to one of six check groups. Each group answers a distinct question about the site. The groups exist because crypto rankings get determined by signals that span the marketing site, the on-chain layer and the AI search layer.
Technical SEO foundation
Core Web Vitals, robots.txt, XML sitemap, canonical, render-blocking resources, mobile viewport, JavaScript rendering, internal linking. Standard SEO foundation, applied to crypto context.
Schema and structured data
FinancialProduct on token pages, CryptoExchange on platforms, SoftwareApplication for wallets and SDKs, Organization plus Person for E-E-A-T. Schema correctness validated against type-specific rules.
YMYL and trust signals
Audit firm citations (Trail of Bits, OpenZeppelin, CertiK), founder identity coverage, regulatory disclosure, domain WHOIS trust, GitHub repository activity. Trust layer specific to financial sites.
AI Visibility (AEO)
4-factor citation model: schema correctness, factual density, AI bot access, authority citations. Citation rate testing across ChatGPT, Perplexity and Claude with category-tuned prompt sets.
Authority and backlinks
Web3 Backlink Toxicity Rubric. Tier 1 sources (DefiLlama, CoinGecko, Etherscan, audit firms) weighted higher than generic high-DA sources for crypto. PBN cluster detection.
On-chain signals
TVL, holder distribution, contract security, GitHub activity, audit firm coverage. On-chain signals provide trust verification that traditional SEO cannot capture.
Why 6 groups, not 8
The 8 audit dimensions in the Crypto SEO Guide describe what the audit covers. The 6 check groups describe how the analyzers are organized internally. Keyword Intent and Competitor Analysis are workflows that span multiple check groups, not standalone groups.
The 23 analyzers in detail
Each analyzer runs as a discrete check with named inputs, calibrated thresholds and explicit output. The analyzers below are listed with their check group prefix and a one-line description.
Core Web Vitals
Technical · LCP, CLS, INP thresholds
Measures LCP under 2.5s, CLS under 0.1, INP under 200ms. Sources from PageSpeed Insights API. Tests mobile and desktop separately.
Robots and crawl access
Technical · robots.txt validation
Validates robots.txt syntax, checks for accidental Disallow rules on key paths, confirms 7 AI bot allow rules (GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot, Google-Extended, Applebot-Extended).
Canonical and indexability
Technical · canonical correctness
Tests canonical tag presence, self-referencing canonicals on key templates, parameter handling, multi-chain canonical patterns. Catches the most common crypto site indexation issue.
JavaScript render coverage
Technical · server-render detection
Tests whether key facts (tokenomics, team info, audit citations) are server-rendered or behind JavaScript. 69% of AI crawlers cannot execute JS so this is a citation-blocking check.
XML sitemap and internal linking
Technical · structure validation
Validates sitemap presence and freshness, identifies orphan pages, checks pillar-cluster topology, validates breadcrumb consistency. Crawl architecture audit.
FinancialProduct schema validator
Schema · token page schema
Detects token information pages, validates FinancialProduct usage, checks for required and recommended properties. Catches the most common crypto SEO mistake (generic Product schema on tokens).
CryptoExchange schema validator
Schema · exchange page schema
Detects exchange platform pages, validates CryptoExchange usage, checks supportedAssets array, tradingFee, address, hasOfferCatalog patterns.
SoftwareApplication validator
Schema · wallet/SDK schema
Detects wallet, SDK and protocol client pages, validates SoftwareApplication with correct applicationCategory (FinanceApplication for wallets, DeveloperApplication for SDKs).
Organization and Person schema
Schema · E-E-A-T signal validation
Validates Organization on company pages, Person on team pages, sameAs array linking to LinkedIn, GitHub and Twitter. Critical for YMYL trust signals.
Audit firm citation detection
YMYL · trust signal
Detects citations to Trail of Bits, OpenZeppelin, CertiK, Quantstamp and other audit firms. Validates audit dates, firm names, PDF link integrity.
Founder identity coverage
YMYL · E-E-A-T author signals
Checks for named founders with Person schema, sameAs links to verified social profiles, professional history. Anonymous-founder projects score lower on YMYL trust.
Regulatory disclosure
YMYL · compliance signal
Detects regulatory status disclosure (FinCEN MSB, MiCAR, BitLicense, FCA), geographic restrictions, "not investment advice" language on relevant pages.
Domain WHOIS trust
YMYL · domain age and history
Checks domain registration date, WHOIS privacy patterns, historical ownership changes. Newer domains score lower on YMYL trust by default.
AEO schema readiness
AI Visibility · factor 01
Tests whether key facts are exposed in JSON-LD that AI extractors can parse reliably. Cross-references with B01-B04 schema validators.
Factual density scorer
AI Visibility · factor 02
Measures named entities, numbers and dates per 100 words. Below 3 per 100 the page reads as marketing rather than reference, which AI engines weight lower.
AI bot access tester
AI Visibility · factor 03
Tests crawl access for 7 AI bots using actual user-agent strings. Catches CDN-level filtering that robots.txt does not show. Edge filtering blocks bots before robots.txt.
Citation rate measurement
AI Visibility · live LLM testing
Runs 30+ category-relevant prompts against ChatGPT, Perplexity and Claude. Records citation rate per LLM, per intent bucket, per competitor cited alongside.
Web3 Backlink Toxicity Rubric
Authority · backlink scoring
Scores backlinks against the open Tier 1/2/3 framework. ~80 Tier 1 crypto authority sources recognized. PBN clusters and toxic neighborhoods flagged for disavow.
Authority source coverage
Authority · DefiLlama, CoinGecko
Detects whether the project is listed on DefiLlama, CoinGecko, CoinMarketCap, Etherscan. These two cover ~60% of crypto AI citation prompts.
PBN cluster detection
Authority · toxicity flags
Detects PBN cluster patterns by hosting analysis, registration overlaps, content thinness signatures and link velocity anomalies. Generates disavow file output.
On-chain trust signals
On-chain · TVL, holders
Pulls TVL from DefiLlama, holder distribution from Etherscan, market cap from CoinGecko. On-chain trust signals weight heavily for crypto AI citation.
GitHub activity scorer
On-chain · repository freshness
Measures public repository activity: commit frequency, named contributor count, recent release tags. AI engines weight active GitHub activity as a "is this project alive" signal.
Smart contract security signals
On-chain · contract trust
Cross-references audit firm reports, checks contract verification status on block explorers, validates ownership renouncement claims, detects honeypot patterns.
Scoring model and weights
Each check group contributes a fixed percentage to the overall site score. Within each group, individual analyzers contribute roughly equal weight. The total adds to 100.
| Check group | Analyzers | Weight | Why this weight |
|---|---|---|---|
| Technical SEO foundation | 5 | 15% | Foundation. Necessary but not crypto-differentiating. |
| Schema and structured data | 4 | 20% | High leverage for AEO. Wrong type kills citation eligibility. |
| YMYL and trust signals | 4 | 15% | YMYL is mandatory for crypto rankings. Anonymous projects score lower. |
| AI Visibility (AEO) | 4 | 20% | Crypto AEO matters more than for any other vertical. |
| Authority and backlinks | 3 | 15% | Web3 Backlink Toxicity Rubric weights crypto authority sources higher. |
| On-chain signals | 3 | 15% | On-chain trust signals are unique to crypto, weight matches AEO. |
| Overall score | 23 | 100% | Weights reviewed quarterly against 53+ live audits. |
Within each group, analyzer-level weights are roughly equal. Schema group weights skew slightly toward FinancialProduct (B01) because it is the highest-impact single fix on most crypto sites. AI Visibility group weights skew toward citation rate measurement (D04) because that is the only direct outcome measurement.
Site type adjustments
The keyword intent classifier (separate from the 23 analyzers) detects site type and applies adjustments. Token information sites get more weight on schema and on-chain signals. Exchange platforms get more weight on YMYL and authority. Wallet sites get more weight on schema and reviews.
Calibration sources
Calibration is the part most audit tools hide. The Crawlux methodology calibrates against three source types: public crypto data APIs, public audit firm reports and internal audit data across live sites.
Source type 01
Public crypto data APIs
On-chain authority signals, market data, contract verification, holder distribution. Calibrates the on-chain check group and authority source coverage analyzer.
- DefiLlama API (TVL data)
- CoinGecko API (market cap, volume)
- Etherscan, Polygonscan, Solscan
- CoinMarketCap (asset listings)
Source type 02
Audit firm public reports
YMYL trust signal calibration. Public audit reports define what "audited and secure" actually means in citable terms for the YMYL group.
- Trail of Bits public reports
- OpenZeppelin Audits
- CertiK Skynet
- Quantstamp public reports
Source type 03
Internal audit calibration
Dimensional weight calibration across 53+ live site audits. Identifies which analyzers correlate with actual ranking and citation outcomes.
- 53+ audited live sites
- Quarterly weight recalibration
- 30+ prompts × 3 LLMs per audit
- Anonymized aggregate metrics
Source type 04
External SEO calibration
Cross-validation against established SEO data sources for technical SEO and backlink dimensions. Used to verify Crawlux scores align with external benchmarks.
- Google PageSpeed Insights API
- DataForSEO (backlinks)
- Schema.org validator
- Google Rich Results Test
What calibration does NOT use
Generic Domain Authority and Domain Rating are intentionally excluded from crypto authority calibration. Generic DA models score DefiLlama and Etherscan as moderate when they are Tier 1 for crypto. Using them would propagate that error into Crawlux scores.
Where it is open vs proprietary
The methodology is open. The execution is proprietary. The line between the two is worth being explicit about.
Open and documented
Anyone can use these
- The 6 check group structure
- The 23 analyzer definitions
- The scoring weights table
- Web3 Backlink Toxicity Rubric tiers
- 4-factor AEO citation model
- Per-analyzer thresholds and rules
- Calibration source list
Proprietary execution
What Crawlux runs
- Internal calibration data across 53+ audits
- Specific weight tunings per category
- Full prompt sets (30+ per category)
- Auto-detection logic for site categorization
- Performance optimizations and caching
- Programmatic LLM citation parsing
- White-label PDF generation pipeline
Other auditors are welcome to use the open methodology, build their own implementations and even compete with Crawlux directly. The published rules are not the moat.
The moat is running 23 analyzers across 30+ AI prompts in 2 minutes, generating a 23-page report and integrating into a quarterly audit workflow at agency scale.
Versioning and review cadence
The methodology gets reviewed on a fixed cadence. Minor revisions adjust weights and thresholds. Major revisions reconsider the analyzer set.
Quarterly minor revisions
90-day cadence
- Weight adjustments based on new audit data
- Threshold updates per analyzer
- New AI bot user-agents added to D03
- Calibration source refreshes
- Per-category weight tuning
Annual major reviews
Once per year
- Reconsider analyzer set (add or deprecate)
- Restructure check groups if needed
- Revise scoring model fundamentals
- Update Web3 Backlink Toxicity Rubric tiers
- Cross-validate against industry benchmarks
Version history is public
Every version change gets logged with the date, the change description and the reasoning. The current methodology is v3, published April 2026. Previous versions are archived for audit reproducibility.
Sites audited in earlier versions can be re-scored in the current version on request. Score changes between versions get flagged with the version delta so users understand what changed and why.
How to challenge a score
Every finding in the audit report includes the analyzer that fired, the rule it failed and the calibration source. If you believe the rule is wrong or the calibration is outdated, the methodology team reviews challenges quarterly.
Identify the analyzer
The audit report lists each finding with its analyzer code (A01 through F03). Reference this code in any methodology challenge so the team knows which rule is being disputed.
Cite the calibration source
Each analyzer's calibration source is documented in section 5 of this page. If you believe the source is outdated or inappropriate for your context, that is the technical basis for a challenge.
Submit via contact
Contact the methodology team via the link on the about page. Include the analyzer code, the calibration concern and (ideally) supporting data or sources. Justified challenges feed into quarterly review.
The most common challenges that have changed weights in past reviews: backlink toxicity scoring for newer DeFi protocols where standard PBN signatures need adjustment, YMYL signals for non-US regulatory frameworks (MiCAR, FCA, MAS, ASIC), GitHub activity scoring for projects using monorepos versus distributed repositories.
What does not get changed
Subjective disagreements ("I do not like getting a low score") are not methodology challenges. Justified challenges include: outdated calibration data, miscategorized site type, inappropriate weighting for specific subverticals.
Common questions
Six questions covering the open methodology philosophy, calibration and review process.
Why publish the methodology openly?
Open methodology is verifiable methodology. Buyers evaluating audit tools should be able to see how scores get calculated, where the calibration data comes from and where the rubric is opinionated versus open. Closed methodology is a black box that makes scores feel arbitrary. The Crawlux methodology is documented openly so other auditors can use it with or without Crawlux. The moat is execution at scale, not the rules.
Where does the calibration data come from?
Three sources. First, public crypto data APIs (DefiLlama, CoinGecko, CoinMarketCap, Etherscan) provide on-chain authority signals. Second, audit firm public reports (Trail of Bits, OpenZeppelin, CertiK) calibrate the YMYL trust dimension. Third, internal audit runs across 53+ live sites provide the dimensional weight calibration. The weights get reviewed quarterly against new data.
How often does the methodology get updated?
Quarterly minor revisions, annual major reviews. Minor revisions adjust weights based on new audit data and changes in AI search behavior. Major reviews reconsider the analyzer set: adding new checks, deprecating outdated checks, restructuring check groups. The current version is v3, published April 2026. Version history is tracked publicly.
Can I challenge a score I disagree with?
Yes. Each finding in the audit report includes the rule it failed and the calibration source. If you believe the rule is wrong or the calibration is outdated, contact the methodology team via the contact link on the about page. Justified challenges feed into the quarterly review cycle. The most common challenges that have changed weights: backlink toxicity for newer DeFi protocols and YMYL signals for non-US regulatory frameworks.
Is the methodology better than Ahrefs or Semrush?
It is different, not strictly better. Ahrefs and Semrush methodologies are excellent for general SEO across news, ecommerce and SaaS verticals. The Crawlux methodology is tuned specifically for crypto and Web3, where authority signals (DefiLlama, audit firms) and schema requirements (FinancialProduct, CryptoExchange) differ from generic SEO. Use the right methodology for the vertical. For non-crypto sites, generic SEO methodologies score better. For crypto sites, the crypto-tuned methodology catches more.
What are the 6 check groups?
Technical SEO foundation (Core Web Vitals, robots.txt, canonical, render-blocking). Schema and structured data (FinancialProduct, CryptoExchange, Person). YMYL and trust signals (audit firm citations, founder identity, regulatory disclosure). AI Visibility (4-factor citation model). Authority and backlinks (Web3 toxicity rubric, PBN detection). On-chain signals (TVL, holders, contract security, GitHub activity). Each group contains 3 to 5 analyzers totaling 23 across the audit.
RUN YOUR FIRST AUDIT FREE
See Crawlux on your own crypto site.
No signup, no credit card. Full Web3-tuned audit report in 60 seconds.
Free first audit · No signup · 60 seconds · Full PDF report