How We Score Cam Sites: 6 Categories, $0 Spend
How we score cam sites: 6 weighted categories, $0 editorial spend, we walk through pricing + checkout up to (never past) pay. Open scoring by Alexandra Joly.
How does bestgirlfriend.ai score cam sites?
We score every covered cam site against six weighted categories: Model Variety (18%), Pricing & Tipping Flow (18%), Broadcast Quality (16%), Payment & Geo Coverage (16%), Privacy & Compliance (16%), UX & Mobile (16%). Those combine into a single 0 to 10 composite. Editorial spend is $0: we walk through pricing and checkout up to submit-payment, and ten broadcast rooms are stopwatch-tested across desktop and mobile.
How we score cam sites looks nothing like how we score chat-based AI companion apps, and that is deliberate. There is no synthetic persona to grade. The broadcasters are real human beings whose work is none of our editorial business. What hurts viewers is opaque token economies, mid-stream buffering, and billing descriptors that surprise nobody who has been there. Six categories, calibrated for exactly that.
Each cam Review carries the full scorecard, the last full retest date, the per-category last-tested dates, and the footnoted sources behind every input. We score in public so that an informed reader, a tester from another site, or a platform whose number we got wrong can all point at the same data and argue it. That is the whole point.
Why is your editorial spend $0 on cam reviews?
Recharging tokens would create a question of whether our score reflects what we paid for or what a brand-new visitor sees. We load pricing and checkout exactly as a visitor would, captures token tiers, payment options, geo indicators, default auto-renewal toggles, and upsell prompts, then halts before submit-payment. The protocol stays infinitely repeatable and conflict-free.
Honestly, this is where most cam-review affiliate sites go quiet. They either claim to have spent money they did not spend, or they spend money in amounts they never disclose. Both paths introduce a conflict-of-interest layer on top of the affiliate commission they are already taking. The $0 protocol resolves both problems. We never paid the platforms we cover. The transparency is intentional and load-bearing, softening it in any localization is forbidden, and removing it would invalidate the score.
What we cannot check from outside (the billing descriptor that lands on a credit card statement 24 hours later, the friction of a refund claim, the auto-renewal that triggers after seven-plus days, internal age-record enforcement quality) is scored from aggregated user reports (≥30 recent Trustpilot or Reddit threads per platform) and footnoted on every affected sub-criterion as not directly tested. Readers see exactly which inputs were directly verified and which lean on user reports. No hiding the line.
I'll also say what most reviewers in this space won't: cam platforms pay well, and pretending the score wasn't influenced by commission economics is the easy lie. We lock scores before commission negotiation. If Chaturbate raises our take tomorrow, the Pricing & Tipping Flow number does not move tomorrow. That is the rule. If it ever breaks, we want readers to catch us.
What are the six cam-site scoring categories?
Six categories weighted into a 0 to 10 composite: Model Variety & Volume 18%, Pricing & Tipping Flow 18%, Broadcast Quality 16%, Payment & Geo Coverage 16%, Privacy & Compliance 16%, UX & Mobile 16%. Each carries its own scoring guide, sub-criteria, data sources, and last-tested date so readers see which numbers are fresh.
| Category | Weight | Primary test method | Primary source |
|---|---|---|---|
| Model Variety & Volume | 18% | Peak-time inventory observation × 3 timezones | Direct screenshots + traffic estimates (secondary) |
| Pricing & Tipping Flow | 18% | pricing + checkout walk ($0 spend) | Direct hands-on check + Trustpilot (secondary) |
| Broadcast Quality | 16% | 10 rooms (5 desktop + 5 mobile-web), 5-min observation | Stopwatch log + Reddit complaints (secondary) |
| Payment & Geo Coverage | 16% | checkout-panel capture + Wayback FAQ diff | Direct hands-on check + user reports (footnoted) |
| Privacy & Compliance | 16% | 2257 + ToS + DMCA capture + regulator-record search | Direct capture + tracker scan |
| UX & Mobile | 16% | Lighthouse mobile audit + filter inventory + store ratings | App Store / Play Store (≥30 reviews, 60 days) |
How do you measure model variety?
We log in (free account where required) at three peak times on the same day (US evening, EU evening, APAC evening) and screenshot live broadcaster counts per top-level tab (Featured, New, Couples, Gay, Trans). Tabs under 50 active broadcasters on a major category are flagged thin. Traffic estimates are the secondary proxy when platform counters look implausible.
The three-timezone protocol matters because cam inventory ebbs and flows across the global broadcaster timezone distribution. A single mid-day snapshot from one geography can hide a thin category or inflate a strong one. Observation on the same day rules out week-over-week noise. It is dull work, but the alternative is letting platforms cherry-pick their own peak hours and call that the inventory.
The category also captures language coverage (counted English-speaking room ratio against total active rooms), gay / straight / trans / couples balance, and any tab thinness that would matter to a reader filtering for what they actually want. Reddit threads with recurring complaints about category population are weighted as secondary signals when the platform's own counters look optimistic.
How do you test broadcast quality?
We open ten rooms per platform (five desktop browser, five mobile web, mixing high-traffic and low-traffic broadcasters) and stopwatch time-to-first-frame, count buffer events during a five-minute observation, log default resolution, and check whether mobile web actually plays video or pushes users into the native app. The category scores against worst-case.
The mobile-web parity check is load-bearing. A substantial chunk of cam traffic is mobile, and platforms that quietly degrade mobile-web playback to push native-app downloads earn lower marks than platforms with true browser parity. Default resolution at first connection, audio quality, and any aggressive autoplay-with-audio behavior are also captured.
Trustpilot and Reddit threads with chronic streaming complaints (≥30 reviews mentioning streaming) are weighted as secondary inputs. Where a platform has shipped an infrastructure migration recently, the category re-tests on the migration announcement rather than waiting for the regular six-month cadence.
How is pricing & tipping flow scored?
We load pricing and checkout exactly as a visitor would and record the token-to-dollar tier table, volume discount curve, payment options panel, geo-availability indicators, default auto-renewal toggle, and upsell prompts on the path to checkout. We stop before submit-payment so editorial spend stays $0.
This is the most-load-bearing category on the page because pricing opacity plus billing surprises is the single most-complained-about issue across cam platforms on every channel I read, Trustpilot, Reddit, App Store reviews, the platforms' own forums. Token rates undisclosed up front, currency surcharges that appear at checkout, "premium membership" upsells that auto-renew without prominent disclosure: all of those land in this category.
Pricing & Tipping Flow also captures the generosity of the free interaction layer (can a logged-out user browse rooms, can a free account chat with broadcasters, what is the minimum recharge), tipping friction (is there a one-click tip jar, is there a pre-set tip menu), and refund / chargeback policy clarity. Last reviewed: 2026-04-28.
How does the $0-spend protocol work?
We walk the discovery layer, open the pricing modal, click through to checkout, and save the rendered pages plus screenshots of the token tier table and payment selectors. We stop before submit-payment. Per platform we re-run every three months for Pricing and every six months for Payment & Geo.
It's a $0-spend pricing check with cam-specific handling per platform. Sites with bot protection get a full hands-on session, with per-platform age-gate handling. Larger sweeps and static pricing pages are checked the same way, just faster.
The protocol is infinitely repeatable across the catalog with zero additional editorial spend. Tier 2 and Tier 3 expansion to Royal Cams, ImLive, CamSoda, Flirt4Free, MyFreeCams, and CAM4 requires no additional budget, only tester time.
What does it mean when you say a score is "not directly tested"?
Sub-criteria we cannot test from outside (post-purchase billing descriptor on the credit card statement, refund-claim friction, auto-renewal triggered after seven-plus days, internal age-record enforcement quality) are scored from aggregated user reports (≥30 recent Trustpilot or Reddit threads) and footnoted on every affected category. The input is real and contributes to the score, but it comes from secondary sources rather than direct verification.
This is a disclosure, not a hedge. The input is real and contributes to the score, but it comes from secondary sources rather than direct verification. Readers see exactly where the line falls. The label and the source list are reproduced in the category footnote on every affected Review, and the not-directly-tested sub-criteria are listed in the version history when the labelling logic changes between releases.
The opposite (sub-criteria where we directly captured the input) carries a "Tested 2026-04-28" footnote with the dated artifact: pricing-capture, Lighthouse audit, broadcaster-count screenshot batch referenced where it can be made public.
How is age verification testing handled?
We capture the 2257 page, ToS, Privacy Policy, and DMCA page with date stamps, then cross-check the Wayback Machine for silent edits. We confirm a named custodian-of-records, a dated statement, and a reachable address. We monitor public regulatory actions (FTC, state AGs, GDPR DPAs) and re-test within seven days of material news.
The discipline is documentation rather than speculation. Where a platform publishes 2257 documentation, we cite it. Where it does not, the category scores accordingly. We do not opine on enforcement quality without a documented regulatory action or a public investigation behind the claim. No vibes-based compliance scoring on a topic this serious.
A third-party tracker scan on the discovery layer captures the breadth of trackers, which contributes to the Privacy & Compliance score alongside the documentation review and the regulator-record search. [Source: webbkoll.dataskydd.net, privacy tracker scanner · verified 2026-05-26] is the load-bearing tertiary tool here.
Does Chaturbate's Texas $675,000 settlement affect its score?
Yes, it is reflected in the Privacy & Compliance category with a footnote linking to the Texas Attorney General docket. Whether it pulls the category to 5/10 versus 7/10 depends on the rest of Chaturbate's compliance footprint, 2257 quality, regulatory record elsewhere, ToS specificity, geo-block enforcement.
Settlements and consent orders are not interpreted as automatic disqualifications. They are weighted alongside the rest of the compliance footprint, and they are weighted heavier for platforms whose corporate identity makes recurrence more likely than for platforms with strong remediation evidence post-settlement. The category footnote on each affected Review names the docket, the date, the settlement amount, and the cited statutory basis, with a link to the regulator's source page where one exists publicly [Source: Texas Attorney General, Chaturbate consent decree (Multi Media LLC, $675,000) · verified 2026-05-26].
Why six categories instead of eight?
AI companion apps live or die on conversation persona, image generation, and voice naturalness. Cam platforms have none of those. Cam platforms live or die on broadcaster count at peak, broadcast quality, and the token economy. Forcing the eight-category AI scoring onto cam would reward platforms for things their users do not evaluate.
The two parallel scoring systems share a parent landing at our methodology overview that explains the two-system architecture. Bridge Versus pages comparing cam vs AI render category by category in narrative form rather than reducing the comparison to a single composite, the products are not commensurable on a unified scale, and pretending otherwise would be the kind of false rigor we built this scoring to avoid.
Why don't you use the AI companion scoring for cam sites?
Different products at the category level: cam has no AI persona to score, no image generator, no voice synthesis; AI companion apps do not broadcast or run a token economy. Two parallel scoring systems share a parent landing at /methodology that explains the why, and bridge Versus pages compare them category by category in narrative form.
The architectural decision is intentional. Readers comparing Stripchat and Candy.ai are not asking "which has the higher composite", they are asking "is a real human better for me than a persistent AI persona, given my budget, my timezone, my privacy posture, and what I am actually trying to get out of either product." Bridge Versus pages are written to that reader. The parent methodology page explains why this is not a workaround for missing coverage but a deliberate design.
Composite scores and what the labels mean
| Composite | Label | Editorial treatment |
|---|---|---|
| 9.0+ | Best in class | Featured in cluster Pillars and listicles. |
| 8.0 – 8.9 | Excellent | Default Tier 1 recommendation. |
| 7.0 – 7.9 | Strong | Recommended with one specific named caveat. |
| 6.0 – 6.9 | Mixed | Recommended only for specific named use cases. |
| 5.0 – 5.9 | Weak | Listed for completeness; not actively recommended. |
| 4.0 – 4.9 | Skip | Explicit "do not recommend" verdict. |
| ≤ 3.9 | Avoid | Listed only when search demand requires; flagged with safety / regulatory warnings. |
How fresh is each score?
Pricing & Tipping Flow re-tests every three months (or on detected price-page change). Model Variety re-tests every three months. Broadcast Quality, Payment & Geo, UX & Mobile re-test every six months. Privacy & Compliance re-tests every six months plus within seven days of any regulatory news.
| Category | Regular cadence | Bypass triggers |
|---|---|---|
| Pricing & Tipping Flow | Every 3 months | Detected price-page change |
| Model Variety & Volume | Every 3 months | Traffic shift >25% MoM |
| Broadcast Quality | Every 6 months | Infrastructure migration announce |
| Payment & Geo Coverage | Every 6 months | Regulatory action or processor change |
| Privacy & Compliance | Every 6 months | Within 7 days of any regulatory news |
| UX & Mobile | Every 6 months | UI overhaul or major app version |
Last reviewed: 2026-04-28.
Can I see raw broadcast tests?
Each Review's scorecard footnotes link to dated public artifacts where they exist, Trustpilot snapshots, Lighthouse audits, App Store rating snapshots, regulatory dockets. Internal artifacts are not republished as raw files, but the editorial team can email a sanitized summary on request to verifiable journalists, researchers, or platforms contesting a score.
Raw artifacts are gated for two reasons. The first is broadcaster privacy: peak-time screenshots can identify individual performers even with redaction, and that is not a risk we take. The second is operational security: the exact way we capture this evidence is part of the testing process we maintain across the catalog. Sanitized summaries cover the underlying numbers without exposing either. Email [email protected] with the URL of the Review and your verifiable affiliation, and you get the summary.
What re-test triggers exist?
Five trigger events bypass the regular cadence: regulatory action announced (state AG, FTC, GDPR DPA), processor or payment partner change, infrastructure migration, UI overhaul or major app version, and traffic shift greater than 25% month over month. Privacy & Compliance re-tests within seven days of any regulatory news.
The trigger framework is mirrored from the AI companion scoring with cam-native additions. The seven-day Privacy & Compliance trigger is the single most-load-bearing safeguard against stale compliance scoring during a period when state-level age-verification statutes (Texas, Utah, Louisiana) and the UK Online Safety Act are actively reshaping the operating environment for every platform we cover [Source: Online Safety Act 2023 (UK), overview · verified 2026-05-26].
How do I report an error?
Email [email protected] with the URL of the affected Review and the specific claim you are contesting, with sources where you have them. Every correction is logged at the top of the affected page for 60 days and noted in the version history if material. Platform contestations follow the same process.
The 60-day correction-log window is deliberate. Readers who came in via search during that window see the corrected page with the correction notice still visible at the top, so they understand what changed and when. Platform contestations get the same treatment, a documented response from a platform we cover is logged the same way a reader correction is logged. We never bury a documented disagreement.
Editorial spend transparency
How affiliate links work on this site
bestgirlfriend.ai earns a commission when a reader signs up to a cam platform through one of our links. Affiliate revenue is the entirety of our business model, and we disclose it on every commercial page per the affiliate disclosure policy [Source: FTC 16 CFR Part 255, Endorsements & Testimonials · verified 2026-05-26]. Three rules govern how affiliate economics interact with editorial:
- Affiliate payouts do not influence scores. When a cam platform raises our commission rate, its composite score does not move. When the same platform's pricing actually improves for the reader (say, a token rate cut), that shows up, but only in the Pricing & Tipping Flow category and only because real users benefit.
- We promote what we'd recommend if we earned nothing. A platform whose composite score is below 5.0 is not promoted, even when the commission is high. A platform we'd never recommend for safety reasons (Privacy & Compliance below 5) is excluded from the catalog regardless of payout.
- CTA freshness is automated. Every CTA on the site shows a "Last verified" date, automatically refreshed against our offer source-of-truth.
Sources
The scoring cites public, dated artifacts on every category. Six load-bearing references:
- [Source: FTC 16 CFR Part 255, Endorsements & Testimonials · verified 2026-05-26], the regulatory basis for our affiliate disclosure rule and our $0-spend transparency posture.
- [Source: Texas Attorney General, Chaturbate consent decree (Multi Media LLC, $675,000) · verified 2026-05-26], the regulatory citation referenced in the Privacy & Compliance footnote on Chaturbate Reviews.
- [Source: Online Safety Act 2023 (UK), overview · verified 2026-05-26], backdrop for the seven-day Privacy & Compliance re-test trigger when UK guidance changes.
- The $0-spend pricing + checkout walk is done by hand exactly as a visitor would, recording pricing tiers without ever submitting payment.
- [Source: webbkoll.dataskydd.net, privacy tracker scanner · verified 2026-05-26], third-party tracker scan on the discovery layer feeding Privacy & Compliance.
- [Source: Wayback Machine, Internet Archive · verified 2026-05-26], silent-edit cross-check across 2257, ToS, Privacy Policy, and DMCA pages.
Cite this page
When citing how we score cam sites in academic, journalistic, or platform contestation contexts:
Joly, A. (2026). How We Score Cam Sites: 6 Categories, $0 Spend. bestgirlfriend.ai. https://bestgirlfriend.ai/methodology/cam-sites
Frequently asked questions
How does bestgirlfriend.ai score cam sites?
We score every covered cam site against six weighted categories (Model Variety (18%), Pricing & Tipping Flow (18%), Broadcast Quality (16%), Payment & Geo Coverage (16%), Privacy & Compliance (16%), UX & Mobile (16%)) combined into a single 0 to 10 composite. Editorial spend is exactly $0. We walk through pricing and checkout up to (never past) submit-payment, and ten broadcast rooms are stopwatch-tested across desktop and mobile.
Why is your editorial spend $0 on cam reviews?
Recharging tokens would create a question of whether our score reflects what we paid for or what a brand-new visitor sees. We load pricing and checkout exactly as a visitor would and record the token tiers, payment options, geo indicators, default auto-renewal toggles, and upsell prompts, then stop before submit-payment. Post-purchase reality is footnoted with its user-report sourcing and flagged as not directly tested.
What are the six cam-site scoring categories?
Six categories weighted into a 0 to 10 composite: Model Variety & Volume 18%, Pricing & Tipping Flow 18%, Broadcast Quality 16%, Payment & Geo Coverage 16%, Privacy & Compliance 16%, UX & Mobile 16%. Each carries its own scoring guide, sub-criteria, data sources, and last-tested date so readers see which numbers are fresh.
How do you measure model variety?
We log in (free account where required) at three peak times on the same observation day (US evening, EU evening, APAC evening) and screenshot the live broadcaster count per top-level tab. Tabs under 50 active broadcasters on a major category are flagged thin. Where a platform exposes a live counter we cite it; otherwise traffic estimates are the secondary proxy.
How do you test broadcast quality?
We open ten rooms per platform (five on desktop browser, five on mobile web) and stopwatch time-to-first-frame, count buffer events during a five-minute observation, log default resolution, and check whether mobile web actually plays video or pushes users into the app. The category scores against the worst-case across the ten rooms.
How is pricing & tipping flow scored?
We load the pricing page and the checkout page exactly as a visitor would and record the token-to-dollar tier table, volume discount curve, payment options panel, geo-availability indicators, default auto-renewal toggle, and any upsell prompts. We stop before submit-payment so editorial spend stays $0. Trustpilot complaints (≥30 recent reviews) are the secondary input.
How does the $0-spend protocol work?
We walk the discovery layer, open the pricing modal, click through to checkout, and save the rendered pages plus screenshots of the token tier table and payment selectors. We stop before submit-payment. Per platform we re-run the protocol every three months for Pricing and every six months for Payment & Geo.
What does it mean when you say a score is 'not directly tested'?
Sub-criteria we cannot test from outside (post-purchase billing descriptor, refund-claim friction, auto-renewal triggers after seven-plus days, internal age-record enforcement quality) are scored from aggregated user reports (≥30 recent reviews) and footnoted on every affected category. The input is real and contributes to the score, but it comes from secondary sources rather than direct verification.
How is age verification testing handled?
We capture the 2257 page, ToS, Privacy Policy, and DMCA page with date stamps, then cross-check the Wayback Machine for silent edits. We confirm a named custodian-of-records, a dated statement, and a reachable address. Internal enforcement quality is not testable from outside, so we monitor public regulatory actions and re-test within seven days of any material announcement.
Does Chaturbate's Texas $675,000 settlement affect its score?
Yes, it is reflected in the Privacy & Compliance category with a footnote linking to the Texas Attorney General docket. Whether it pulls the category to 5/10 versus 7/10 depends on how the settlement compares to the rest of Chaturbate's compliance footprint. Each Review's category footnote shows the contribution of every data point.
Why six categories instead of eight?
AI companion apps live or die on conversation persona, image generation, and voice naturalness. Cam platforms have none of those. Cam platforms live or die on broadcaster count at peak, broadcast quality, and the token economy. Forcing the eight-category AI scoring onto cam would reward platforms for things their users do not evaluate.
Why don't you use the AI companion scoring for cam sites?
Different products at the category level. Two parallel scoring systems share a parent landing at our methodology overview that explains the architecture, and bridge Versus pages compare cam vs AI category by category in narrative form rather than collapsing them into a single composite that would mislead readers.
How fresh is each score?
Pricing & Tipping Flow re-tests every three months. Model Variety re-tests every three months. Broadcast Quality, Payment & Geo, UX & Mobile re-test every six months. Privacy & Compliance re-tests every six months and within seven days of any regulatory news. Each Review hero shows last full retest plus per-category last-tested dates.
Can I see raw broadcast tests?
Each Review's scorecard footnotes link to dated public artifacts where they exist. Internal artifacts (peak-time inventory screenshots, broadcast test logs) are not republished as raw files, but the editorial team can email a sanitized summary on request to verifiable journalists, researchers, or any platform contesting a score.
What re-test triggers exist?
Five trigger events bypass the regular cadence: regulatory action announced, processor or payment partner change, infrastructure migration announced, UI overhaul or major app version, and traffic shift greater than 25% month over month. Privacy & Compliance carries the most aggressive trigger, within seven days of any regulatory news.
How do I report an error?
Email [email protected] with the URL of the affected Review and the specific claim you are contesting, with sources where you have them. Every correction is logged at the top of the affected page for 60 days and noted in the version history if the change is material.
Trust cluster
- Methodology overview, parent landing covering the two-scoring-system architecture
- AI companion methodology, sibling scoring system (8 categories)
- Adult game methodology, sibling scoring system (7 categories, billing-transparency category)
- About bestgirlfriend.ai and Alexandra Joly, masthead, editorial team, corporate identity
- Editorial process, how briefs are commissioned, reviewed, and published
- Affiliate disclosure, FTC 16 CFR Part 255 compliant disclosure (commercial-page footer)