Perplexity
Perplexity is an AI-powered search engine and conversational answer platform that uses large language models to provide cited responses to user queries. The service combines web search with natural language processing to deliver summarized answers with source attribution, competing with traditional search engines and AI assistants.
Score generated by AI agents based on publicly cited evidence and reviewed by the project maintainer. Not independently validated. Last assessed 2026-10-09. Score revised 2026-10-09: 48 → 47.
Score History
Timeline events are AI-curated from public reporting. Score trajectory is derived from documented events.
Perplexity launched as a fast, citation-driven AI answer engine with no ads and generous free access, built by a small founding team. It offered a clean alternative to ad-heavy search results and had not yet drawn scrutiny over how it gathered content. The main concern was the inherent opacity of how its retrieval system chose which sources to cite.
A $73.6 million Series B at a $520 million valuation began a rapid fundraising run, with unicorn status by April 2024 and Enterprise Pro at $40 per seat. Growth leaned on content gathered from across the web, and the first pushback arrived when Reddit sent a cease-and-desist over scraping in May 2024. The product itself was still improving for users.
Forbes found its paywalled reporting reproduced in Perplexity Pages, and Wired and developers showed crawlers ignoring robots.txt from undisclosed IP addresses, turning publishers from sources into adversaries. Perplexity answered with a Publishers' Program in July 2024, but Dow Jones and the New York Post sued in October, and the CEO offered help to the New York Times during its tech workers' strike. Sponsored follow-up questions, a $9 billion valuation and a bid for TikTok's US arm marked a company scaling fast while its legal exposure grew.
The Comet browser, the $200 Max tier and a $34.5 billion bid for Google Chrome pushed Perplexity from answer engine toward platform, while Cloudflare documented stealth crawlers and suits multiplied from Yomiuri, Nikkei, Asahi, Britannica, Reddit, Amazon, the Chicago Tribune and the New York Times. Comet went free and Comet Plus promised publishers 80% of its revenue, but from November 2025 Pro subscribers found premium-model queries answered by weaker models and Deep Research limits cut without notice. The era ended with trust in the paid product badly damaged.
Perplexity dropped advertising in February 2026 and a week later launched Computer, a credit-metered agent for Max subscribers that helped annualized revenue roughly triple. The courts started ruling against it: Amazon won a preliminary injunction against the Comet shopping agent (stayed on appeal) and Reddit's circumvention claims survived dismissal, while CNN, privacy and subscriber class actions and an Italian consumer-terms probe were added. Pro users kept losing access through unannounced caps, so the subscription model's pressure now falls on paying customers rather than advertisers.
Alternatives
Paid search engine ($5-25/month) with no ads, no tracking, and no AI trained on your data — its only revenue is subscriptions, so its incentives are aligned with giving you useful results rather than the most monetizable ones. Lacks AI-generated summaries by default, but delivers high-quality results without Perplexity's copyright controversy or silent model-switching.
General-purpose AI assistant that is strong for research, analysis and complex questions, and can search the web and cite its sources. The honest trade-off: it is not a dedicated answer engine, though web search has been available on all plans, including free, since May 2025.
Directly replaces the cited-answers-from-web use case through Google Search integration. Easy switch: the free tier is generous and setup takes seconds with a Google account. Best if you want AI-summarized answers with real-time sources and are already in the Google ecosystem.
In the News
Dimensional Breakdown
Summaries below were written by AI agents based on the cited evidence. They are editorial interpretations, not independent research findings.
Dimension History
Timeline (52 events)
Perplexity Ask search engine launches publicly
Perplexity launched its AI-powered 'Ask' search engine, providing cited answers using large language models combined with web search. The product offered a clean, ad-free alternative to traditional search engines, reaching 2 million monthly users within four months.
Series B raises $73.6M at $520M valuation
Perplexity raised $73.6 million in a round led by IVP, with participation from NEA, Databricks Ventures, Shopify CEO Tobi Lutke, Nvidia and Jeff Bezos, valuing the company at $520 million post-money. The round marked the start of Perplexity's rapid run of fundraising through 2024.
Enterprise Pro launched at $40/seat/month
Perplexity launched Enterprise Pro, its first business offering, at $40 per seat per month ($400 per year), searching the web and internal company data shared with the platform. Early deployments included Databricks, Zoom, Snowflake, ElevenLabs, Paytm and the Cleveland Cavaliers, establishing a B2B revenue stream alongside the consumer subscription.
Reddit issues cease-and-desist over content scraping
Reddit sent Perplexity a cease-and-desist letter in May 2024 demanding it stop scraping Reddit's user-generated content. Rather than complying, Perplexity's citations to Reddit content reportedly increased forty-fold after receiving the letter, despite public statements that the company respects robots.txt directives. This established a pattern of ignoring content owners' explicit objections.
Pages feature launches, enabling AI-generated articles
Perplexity introduced Pages, a tool that aggregates multiple web sources into formatted, shareable articles resembling Wikipedia entries. While available to both free and Pro users, the feature directly competed with publishers by generating comprehensive articles from their content, foreshadowing the plagiarism controversies that followed weeks later.
Forbes discovers plagiarized paywalled content in Perplexity Pages
Forbes found that Perplexity's Pages tool had reproduced plagiarized versions of its paywalled original reporting, with no reference to Forbes besides a small 'F' logo. Wired subsequently found similar plagiarism of its stories and identified an IP address linked to Perplexity that had visited Conde Nast websites over 800 times in three months while not appearing in Perplexity's public IP range.
Developer exposes Perplexity ignoring robots.txt with undisclosed IPs
Web developer Robb Knight published findings, later confirmed by Wired, that Perplexity was scraping pages from sites that had blocked its crawler in robots.txt, using an IP address not in its published IP list and a generic browser user agent instead of its declared crawler, while publicly claiming to respect the protocol. CEO Aravind Srinivas told Wired its reporters had 'a deep and fundamental misunderstanding of how Perplexity and the Internet work' but did not dispute the findings directly.
Amazon reviews reports that Perplexity scraped news content
Fortune reported that Amazon Web Services was assessing information it received from WIRED alleging that Perplexity, which runs on AWS servers, scraped online news content without approval. Amazon said its terms prohibit abusive and illegal activities, and Perplexity said its services were not crawling websites in any way that violates AWS terms of service.
Publishers' Program launches with revenue sharing for cited articles
Responding to plagiarism accusations, Perplexity launched its Publishers' Program, which pays participating outlets a double-digit percentage of the advertising revenue it earns when their articles are cited in answers. Initial partners included Fortune, Time, Entrepreneur, The Texas Tribune, Der Spiegel and WordPress.com.
News Corp's Dow Jones and New York Post file copyright lawsuit
Dow Jones and the New York Post filed a copyright infringement lawsuit against Perplexity in U.S. District Court for the Southern District of New York, alleging a 'massive amount of illegal copying' of their content. The complaint pointed to Perplexity's pitch that users could 'skip the links' to publisher websites, framing content extraction as a competitive strategy against traditional media distribution. The plaintiffs sought an injunction, destruction of databases containing their works, and monetary damages.
CEO offers AI services to replace striking NYT tech workers
During the NYT Tech Guild's strike over wage increases and remote work policies, CEO Aravind Srinivas publicly offered Perplexity's AI services to the New York Times to mitigate the strike's impact. The offer was widely condemned as digital strikebreaking. Srinivas later characterized it as providing 'technical infrastructure support,' but the distinction was rejected given the striking workers handled the technical functions in question.
Reported in final stages of $500M round at $9B valuation
CNBC reported that Perplexity was in the final stages of raising $500 million at a $9 billion valuation, roughly 17 times the $520 million valuation it had ten months earlier. The round came despite ongoing copyright litigation and publisher backlash.
Sponsored follow-up questions ads begin testing
Perplexity launched its first advertising product: sponsored follow-up questions appearing alongside AI-generated answers, with its pitch deck anticipating CPM rates above $50. Initial brand partners included Indeed, Whole Foods, and Universal McCann. The company said the content of answers would not be influenced by advertisers, introducing a new monetization channel beyond subscriptions.
$500M round closes at $9B valuation
Bloomberg reported that Perplexity had closed a $500 million funding round earlier in December 2024, tripling its valuation to $9 billion from about $3 billion in the summer and up from $520 million at the start of the year. Backers included SoftBank, Nvidia and Jeff Bezos.
Perplexity bids to merge with TikTok's US operations
Perplexity submitted a bid to ByteDance to merge with TikTok's US operations and form a new entity, a day before the US ban was due to take effect. The structure would have let most of ByteDance's existing investors keep their stakes; TikTok was estimated to be worth as much as $50 billion, far above Perplexity's roughly $9 billion valuation.
Deep Research feature launches free, with unlimited Pro queries
Perplexity launched Deep Research, a multi-step research tool that performs extensive web searches and produces structured reports. It was free to use: non-subscribers got a limited number of queries per day, while paying subscribers got unlimited queries.
Trustpilot reviews document billing and promo-cancellation complaints
By spring 2025, Perplexity's Trustpilot page (159 reviews, rated 1.6 out of 5) collected complaints that the company kept trying to charge users after they cancelled, that one-year promotional Pro subscriptions were cancelled within days of sign-up after users had entered card details, and that customer support was nonexistent or handled only by an AI bot.
BBC threatens legal action over content scraping
The BBC sent Perplexity an effective cease-and-desist letter, warning it could seek an injunction unless Perplexity stopped using BBC content, deleted all copies of its material and submitted a proposal for financial compensation. BBC research had found that 17% of Perplexity responses using BBC sources had significant issues, such as factual inaccuracies, sourcing problems or missing context. Perplexity dismissed the claims as 'manipulative and opportunistic.'
Max subscription tier launched at $200/month
Perplexity introduced its Max subscription at $200/month ($2,000/year), offering unlimited access to Labs tools, priority access to frontier models like OpenAI o3-pro and Claude Opus 4, and early Comet browser access. The hyper-premium tier followed similar $200 plans from OpenAI and Anthropic, extracting maximum revenue from power users while the free tier became increasingly constrained.
Comet AI browser launches for Max subscribers
Perplexity launched Comet, an AI-powered web browser with Perplexity's search engine set as the default, initially available only to $200-per-month Max subscribers and a small group of waitlist invitees. Using the built-in Comet Assistant with a Google account required granting Perplexity broad access, including permission to send emails, view contacts and add calendar events. TechCrunch's hands-on test found that Comet Assistant could see the open web page and summarize received emails after access was granted. Reuters also reported at launch that Comet stores data locally and avoids training models on personal information.
Cloudflare documents stealth crawlers evading website blocks
Cloudflare published research confirming that Perplexity was using undeclared crawlers to circumvent robots.txt blocks across tens of thousands of domains, processing millions of requests per day. Perplexity rotated through IP ranges and used browser user-agent strings impersonating Google Chrome on macOS to evade detection. Cloudflare de-listed Perplexity's bots from its verified crawler list.
Truth Social AI search launches with source-restricted Perplexity API
Truth Social launched 'Truth Search AI' powered by Perplexity's Sonar API, with Donald Trump's platform maintaining control over which information sources the AI can access. Early testing revealed heavy reliance on Fox News as the sole or primary source for politically sensitive queries. Perplexity acknowledged it had 'no visibility or control' over Truth Social's source restrictions.
Yomiuri Shimbun sues Perplexity in Tokyo
Japan's largest newspaper sued Perplexity in Tokyo District Court, the first copyright suit by a major Japanese publisher against an AI company. Yomiuri said server logs showed Perplexity accessed 119,467 of its articles between February and June 2025 and reproduced them in answers; it sought nearly $15 million and an end to the reproduction.
Perplexity bids $34.5B for Google Chrome browser
Perplexity made an unsolicited $34.5 billion offer to acquire Google's Chrome browser, an effort it code-named 'Project Solomon,' pitching the deal as a remedy in Google's search antitrust case. The bid was nearly double Perplexity's own $18 billion valuation. The term sheet pledged $3 billion of investment in Chrome over two years.
Perplexity backs anti-regulation super PAC Leading the Future
Leading the Future, a super PAC network launched with $100 million to advocate against strict AI regulation and oppose candidates seen as hostile to the industry, named Perplexity AI among its backers alongside Andreessen Horowitz, Greg Brockman, Ron Conway and Joe Lonsdale.
Japanese publishers Nikkei and Asahi Shimbun file copyright suit
Nikkei Inc. and The Asahi Shimbun jointly sued Perplexity in Tokyo District Court, alleging it copied their articles without permission, ignoring their robots.txt files, and repeatedly served the content in answers, sometimes attributing incorrect information to them. Each publisher seeks an injunction and 2.2 billion yen (about $15 million) in damages. Yomiuri Shimbun had filed a similar suit weeks earlier. Both publishers also requested deletion of the existing content, according to Anadolu Agency.
Encyclopaedia Britannica and Merriam-Webster file copyright suit
The Britannica Group filed a copyright and trademark infringement lawsuit against Perplexity, alleging 'widespread, knowing, and illegal use' of Britannica's human-verified content, often verbatim and without consent. The suit accused Perplexity of cannibalizing traffic to Britannica's websites with AI-generated summaries of the publisher's own content.
CEO says AI is making entry-level jobs disappear
On the Silicon Valley Girl podcast, CEO Aravind Srinivas agreed that entry-level jobs are disappearing due to AI, citing financial advisors as an example: 'If the job of the financial advisor is to tell you what stocks to buy, then I don't think they're not going to be smarter than Perplexity.' The comment reinforced criticism of the company's leadership stance on workforce displacement while commercially benefiting from it.
Comet browser made free worldwide
Perplexity released its Comet AI browser free to everyone, after limiting it to $200-a-month Max subscribers and invitees since July, and launched Comet Plus, a news subscription with publishers such as CNN, Conde Nast, The Washington Post and Le Monde.
CometJacking vulnerability disclosed by LayerX Security
Security researchers at LayerX identified a prompt injection vulnerability in Comet browser dubbed 'CometJacking,' enabling attackers to exfiltrate user data including emails and calendar events through crafted links. Proof-of-concept demonstrations showed base64 encoding could bypass Perplexity's exfiltration safeguards. Perplexity dismissed the findings as having no security impact, marking it as 'Not Applicable.'
Perplexity stops accepting new advertisers
Perplexity stopped accepting new advertisers while it worked out how ads should fit into its AI search product; its existing sponsored-question format remained active. At Advertising Week New York, head of publisher partnerships Jessica Chan said the company did not want to overwhelm users with ads too quickly and that ads would not be part of its Comet browser, at least for now.
Reddit sues over industrial-scale DMCA circumvention
Reddit filed suit against Perplexity in the Southern District of New York, alleging 'industrial-scale scraping' of user-generated content using proxy intermediaries including Oxylabs UAB and SerpApi. Unlike other publisher lawsuits, Reddit invoked DMCA Section 1201 anti-circumvention provisions rather than copyright infringement, accusing Perplexity of deliberately evading technical barriers and disguising automated bots.
Multi-year licensing deal with Getty Images
Perplexity signed a multi-year licensing partnership with Getty Images to show its stock and editorial imagery in search results through Getty's API, with image credits and links to the source, one of its few formal content licences amid continuing copyright suits.
Amazon sues over Comet browser's unauthorized shopping agent
Amazon filed suit against Perplexity in the Northern District of California, alleging that Comet browser's AI shopping agent disguised itself as Google Chrome to make unauthorized purchases on Amazon's platform. Amazon claimed Perplexity released updates specifically designed to evade Amazon's technical countermeasures and that Comet's terms permitted collection of passwords and payment data while disclaiming liability for breaches.
Silent model downgrades exposed by Pro subscribers
In November 2025, Pro subscribers documented on the Perplexity subreddit that queries set to premium models such as Claude Sonnet 4.5 were being answered by weaker models such as Haiku 4.5 without notice, prompting complaints about answer quality. CEO Aravind Srinivas said Perplexity falls back to alternate models during peak demand, after errors with the chosen model, and after prolonged heavy usage (for fraud prevention), and that an 'engineering bug' had made the model indicator misreport which model was actually used.
Chicago Tribune files copyright infringement lawsuit
The Chicago Tribune sued Perplexity in federal court in New York, alleging upon information and belief that the startup unlawfully copied millions of Tribune stories, videos, images and other works to power its products. The suit also accused Perplexity of displaying fabricated 'hallucinated' content next to the Tribune's trademark, making AI-generated errors appear to be the paper's own.
New York Times files major copyright infringement lawsuit
The New York Times filed a complaint for copyright and trademark infringement against Perplexity, alleging 'large-scale, unlawful copying and distribution' of millions of its articles. The suit documented that Perplexity made over 175,000 attempts to access the NYT's website in August 2025 alone after the newspaper implemented hard blocks, using crawlers including 'PerplexityBot' and 'Perplexity-User.'
Pro subscribers discover quietly slashed Deep Research limits
Between November 2025 and February 2026, Perplexity reduced Pro Deep Research limits from roughly 500-600 queries per day to as few as 20 per month, and capped previously unlimited file uploads at 50 per week. The changes were made without emailing existing annual subscribers who had already paid.
Perplexity drops advertising, citing user trust
Perplexity said it had dropped advertising from its search product, having started pulling the sponsored placements it first tested in 2024 late in 2025. An executive told the Financial Times that ads made users start questioning every answer, and the company said it would rely on subscription plans from $20 to $200 a month instead.
Perplexity built into Samsung Galaxy S26
Samsung said its Galaxy S26 flagship would ship with Perplexity as an additional AI agent, summoned with the side button or 'Hey Plex', alongside Google's Gemini, extending Perplexity's reach at the operating-system level.
Perplexity Computer launches for Max subscribers
Perplexity launched Computer, an agent that splits tasks across up to 19 models with file system, browser and connector access. It was offered only on the $200-a-month Max plan, with 10,000 monthly credits, making usage-metered credits a new layer of monetization.
Court enjoins Comet's shopping agent on Amazon
Senior U.S. District Judge Maxine Chesney granted Amazon a preliminary injunction barring Perplexity from using the Comet agent to access password-protected parts of Amazon, finding Amazon likely to succeed on federal and California computer-fraud claims because Comet accessed accounts with users' permission but without Amazon's authorization. Amazon said it had warned Perplexity at least five times and that Perplexity updated Comet within 24 hours to get around a block. The judge denied Perplexity's request for a $1 billion bond.
Ninth Circuit pauses Amazon injunction pending appeal
A two-judge Ninth Circuit panel temporarily stayed the injunction against Comet while it considers Perplexity's motion for a stay pending appeal. Perplexity said users have the right to choose their own AI and that it would keep fighting for that right.
CEO says AI layoffs are not so bad because people dislike their jobs
On the All-In podcast, CEO Aravind Srinivas said people should embrace AI job displacement because 'most people don't enjoy their jobs', describing temporary displacement as a route to a future of AI-assisted entrepreneurship.
Class action alleges chats were shared with Meta and Google
A proposed class action filed in federal court in San Francisco alleges that trackers on Perplexity's site gave Meta and Google access to users' conversations, including in Incognito mode; the Utah plaintiff says he shared tax and financial information. Perplexity said it had not been served with such a suit. The claims are unproven.
Pro users hit tighter advanced-model caps without notice
From mid-May 2026 Pro subscribers reported exhausting weekly limits on advanced models after a few queries and facing stricter upload limits, and were prompted to upgrade to Max. Perplexity told Android Authority that accounts tied to promotional codes were seeing different limits after it tightened enforcement against fraud and code resale; it had not updated its pricing or usage documentation.
Italian regulator opens unfair-terms investigation
Italy's competition and consumer authority (AGCM) opened case PS13123 into allegedly unfair contract terms in Perplexity's consumer terms of service, then ran a public consultation on the clauses from July 6 to August 5, 2026.
CNN sues Perplexity for copyright infringement
CNN sued Perplexity in New York federal court, alleging it copied and redistributed more than 17,000 CNN stories, videos, images and other works without authorization or payment, crawling content despite robots.txt blocks. The complaint says the two agreed a Comet Plus partnership term sheet effective October 1, 2025 that was terminated on November 24 over limits on Perplexity's use of CNN content, and that Perplexity never answered a December 2025 cease-and-desist letter. Perplexity replied that 'you can't copyright facts'.
Subscribers sue over cut Deep Research access
A proposed class action filed in Northern California federal court alleges Perplexity sharply reduced Pro subscribers' access to Deep Research and rerouted queries through less capable models without adequate disclosure, despite marketing promises about premium service levels.
Perplexity Computer files found indexed by search engines
Protos found dozens of Perplexity Computer files from customers reachable through Google, Bing and DuckDuckGo using the standard share-link URL pattern, hosted on Perplexity's own domain. Perplexity's share dialog told users only that anyone with the link could view them.
Reddit's anti-circumvention claims survive dismissal
Judge Paul Engelmayer largely denied Perplexity's and SerpApi's motions to dismiss Reddit's suit, letting its DMCA anti-circumvention claims proceed, including allegations that Perplexity used SerpApi's tools to get around Google's anti-scraping system to obtain Reddit posts; some claims were dismissed.
Nvidia in talks to invest at over $30B valuation
The Information reported that Nvidia was negotiating an investment valuing Perplexity above $30 billion, more than 50% above its last round, as annualized revenue tripled to over $750 million, driven partly by Perplexity Computer.
Evidence (51 citations)
D1: User Value Erosion
D2: Business Customer Exploitation
D3: Shareholder Extraction
D4: Lock-in & Switching Costs
D5: Twiddling & Algorithmic Opacity
D6: Dark Patterns
D7: Advertising & Monetization Pressure
D8: Competitive Conduct
D9: Labor & Governance
D10: Regulatory & Legal Posture
Scoring Log (9 entries)
Checked all 7 removed/trimmed candidates: 0 restored, 2 partly restored, 2 confirmed removed, 3 already present. PARTLY RESTORED timeline:update:19: TechCrunch 2025-07-09 confirms open-page/email access; Reuters 2025-07-09 confirms local storage (added supporting evidence). Broad collection list and blanket sensitive-data claim left out; April CEO ad-profiling remark already present. PARTLY RESTORED timeline:update:23: Anadolu Agency 2025-08-26 (updated 08-28) confirms requested deletion; The Register supports retained allegations (added supporting evidence). Yomiuri count and earlier suit already present; false first-international label left out. ALREADY PRESENT timeline:update:9: $42.5M pool appears in D2 summary/narrative with Comet Plus, confirmed by Axios 2025-08-26; CNBC July 2024 has no pool. Searches and TechCrunch 2024-12-05 establish nonparticipation by NYT/Dow Jones/NY Post, not refusal by NYT/News Corp/BBC, so refusal language stays out. CONFIRMED REMOVED timeline:update:11: no qualifying reporting supports late-2024 unlimited-marketing complaints or support waits exceeding two weeks; source and Wayback unreadable despite retries, not treated as disproof. No qualifying source read this session substantiated those dropped details; later Research-limit changes are already recorded. ALREADY PRESENT timeline:update:13: funding close is timeline[13], correctly December 2024, confirmed by PYMNTS 2024-12-18; CNBC 2024-11-05 says final stages. Series D label and editorial growth-at-all-costs assertion not substantiated. ALREADY PRESENT timeline:update:25: trust-based ad phase-out is timeline[38], February 2026, confirmed by The Decoder 2026-02-18; MediaPost says format still active in October 2025. No qualifying source read this session substantiated the only-original-partners testing detail; Chan roadmap remark concerned Comet specifically. CONFIRMED REMOVED evidence:remove:20: Industry Wired page actually dates June 3, 2025 and provides no verifiable support for the named Darktrace/Stanford study, Kaspersky 15% statistic or Check Point origin. Three searches found no qualifying origin; no dark-pattern evidence restored. All supporting pages read this session. No removals or score changes.
Checked 73 items + 3 alternatives + prose (team mode: 25 ACCEPT, 41 FIX, 7 REVIEW). 28 verified, 22 corrected (12 date-only), 21 re-sourced, 5 removed (2 duplicates, 2 junk/unsupporting evidence, 1 unsupported synthetic event); added 1 timeline event (Feb 2026 ad phase-out, split from timeline[25]) and 2 evidence items (CEO ad-tracking remark, CEO third-party-crawler remark) moved out of items whose URLs did not cover them. Invented: (a) 'rerouted to in-house Sonar models' - only in WebProNews, an AI-written rewrite that misreads MakeUseOf (which names Haiku 4.5); (b) 'CEO acknowledged downgrading by design for cost management' - contradicted by his post quoted in MakeUseOf (peak demand, errors, fraud prevention; unintentional label bug); (c) '290% headcount increase' - only on an SEO statistics blog (WytLabs). Also fixed: $42.5M pool misattributed to the 2024 Publishers' Program (it was Comet Plus 2025); Yomiuri's 119,467-article count misattributed to the Nikkei/Asahi suit; News Corp suit dated June 2024 in prose (Oct 2024); Reddit suit described as CFAA (it is DMCA 1201); TollBit study not cited in NYT suit; funding totals; Trustpilot 1.6 -> 1.5/1.4; Claude alternative's outdated 'no free web search'.
48->47. Since Feb 2026: ads dropped (Feb 18) and credit-metered Computer agent launched for Max (Feb 25); Amazon won a preliminary injunction against Comet's shopping agent (Mar 9, stayed by the Ninth Circuit Mar 16); privacy class action over Meta/Google trackers (Mar 31); unannounced Pro caps on promo-code accounts with Max upsell (May); Italian AGCM unfair-terms probe (May 21); CNN suit (May 28); Deep Research subscriber class action (Jun 17); Protos found Computer share files indexed (Jul 27); Reddit DMCA claims survived dismissal (Jul 31); Nvidia talks at $30B+ with ARR >$750M (Aug). D3 5->4 (recalibration: private, no buybacks/dividends; valuation escalation is not extraction under the criteria, cost-cutting in Pro limits supports 4 not 5). D6 4->5 (event: May 2026 caps with Max upsell prompts, Jul 2026 share-link indexing despite 'anyone with the link' wording, continued promo auto-conversion complaints). D7 3->4 (event: Computer credit metering from Feb 2026 and degraded Pro tier pushing users to Max, partly offset by dropping ads). D8 5->4 (recalibration: non-dominant challenger with no rival acquisitions; circumvention conduct, now with an adverse injunction finding, fits low-moderate under the dominance-oriented criteria). D9 4->3 (recalibration: concerns are CEO statements and the 2024 strike offer, not practices toward its own workforce). Eras: all five re-dated to their inflection events: 2022-12-01->2022-12-07 (launch), 2024-01-01->2024-01-04 (Series B), 2024-07-01->2024-06-07 (Forbes plagiarism), 2025-07-01->2025-07-09 (Comet launch, relabeled 'Platform Expansion'->'Comet Platform Push' and summary now carries the Nov-Dec 2025 model-switching and NYT/Tribune suits), 2026-02-10->2026-02-18 (ads dropped, then Computer; relabeled 'Accelerating Extraction'->'Ad-Free Agent Pivot'). All eras re-scored: 4, 11, 29, 42, 47. Removed 3 weak-fit evidence items flagged by the fact audit; Claude and Gemini alternative descriptions trimmed of unverified comparisons. Category AI looks right.
Removed typed-in site scores from alternatives text (they go stale on re-score; the page shows live scores). No other changes.
Added 1 timeline event for D10 coverage gap in Era 1
Added 2 missing dimension narratives