ElevenLabs Review 2026: A $22B Voice AI Company That Shipped a New Flagship Model the Day We Checked
ElevenLabs started in 2022 as two Polish founders' frustration with badly-dubbed movies. By this week it's a voice-AI company reportedly marked at $22 billion in an in-progress secondary share sale -- roughly double the $11 billion valuation it raised at just seven months earlier -- pacing toward $600 million in annual revenue. On the exact day this review was verified, ElevenLabs shipped Eleven v4 and v4 Turbo, its most capable voice models yet, already ranked #1 on an independent voice-quality leaderboard. None of that is the whole story: the same company is currently a defendant in a pending federal lawsuit over how its voice models were trained, its real-world review record is genuinely split once you look past the headline average, and its credit-based billing system is the single most common source of user complaints. This review checks what's actually true today, against ElevenLabs' own site, its official blog, independent financial and legal reporting, and real user sentiment -- not the pitch.
| Ownership | Independent, founder-led since 2022 |
| Free tier | Real -- 10,000 credits/mo |
| Cheapest paid plan | $6/mo (Starter, monthly) |
| ARR (Sep 2026, CEO-stated pacing) | ~$600M, up from $350M end-2025 |
| Trustpilot record | 3.0/5, 1,165 reviews -- billing-driven complaints |
The voice quality genuinely is the best most people have heard. The billing system and the legal history are the two things worth reading before you commit.
"ElevenLabs' real differentiator isn't just that its voices sound human -- competitors are closing that gap -- it's that the company keeps shipping a genuinely broader platform around that core model faster than most rivals can match it: voice agents with a native MCP server and n8n integration, a music generator, a medical-grade transcription model, and, on the exact day I checked this, a new flagship voice model that's already ranked #1 on an independent benchmark. That pace is why ElevenLabs went from a $3.3 billion valuation in January 2025 to a reported $22 billion by September 2026. But two things sit right next to that growth story: a Trustpilot record (3.0/5 across 1,165 reviews) where the complaints are overwhelmingly about the credit-based billing system rather than voice quality, and a currently pending federal lawsuit -- filed under Illinois' biometric privacy law in May 2026 -- alleging its models were trained on journalists' and voice actors' recordings without consent. Both are true about the same company, in the same year it also settled an earlier, similar lawsuit confidentially."
None of that erases what ElevenLabs does well: the free tier is real and usable, the voice quality genuinely holds up against everything else in the category, and the company is clearly shipping product, not coasting on its valuation headlines. It does mean I'd read the credit/downgrade terms before committing to a plan, budget for the fact that compliance features (SOC 2, HIPAA, Zero Retention Mode) are gated to the Enterprise tier, and treat both the loudest praise and the loudest complaints as real signals about the same product.
One voice engine, spread across eight products -- text-to-speech, agents, music, transcription, dubbing, and a consumer reading app.
ElevenLabs started as a single text-to-speech API and has grown into a full voice-AI platform. As of this check, the current line-up is: Eleven v4 / v4 Turbo (flagship text-to-speech and voice cloning, launched Sep 28, 2026), ElevenAgents (Conversational AI / voice agents deployable on phone, chat, email, and WhatsApp), Eleven Music (natural-language music generation), Scribe v2 and Scribe v2 Medical (speech-to-text, including a clinical/medical-audio specialist model), Studio (a multi-track AI audio editor), Dubbing v2 (translation that preserves emotional performance), Sound Effects (custom SFX/ambient audio), and ElevenReader, a consumer audiobook app with 200,000+ human-narrated titles at $11/month, positioned directly against Audible.
Worth flagging up front: unlike several products JAVIS has reviewed this year, ElevenLabs has not been acquired and has no parent company -- it's still run by its original 2022 co-founders. The company's own growth is the story here, not a change of ownership.
A reported $22B mark, four years after founding -- but that number needs one important caveat most coverage skips.
Backers' mark on an in-progress secondary tender offer, not a new primary raise.
$500M Series D, closed Feb 4, 2026, led by Sequoia.
Pacing figure; up from $350M at end of 2025.
Across 50+ countries, per ElevenLabs' own Sep 2026 update.
| Date | Event | Result |
|---|---|---|
| 2022 | Piotr Dabkowski & Mati Staniszewski found ElevenLabs | Bootstrapped idea: better-than-dubbing AI voice |
| 2024-01 | $80M Series B | $1.1B valuation -- unicorn status |
| 2025-01 | $180M Series C | $3.3B valuation |
| 2025-09 | First $100M employee tender offer | $6.6B valuation |
| 2026-02-04 | $500M Series D closes | $11B valuation, led by Sequoia Capital |
| 2026-05 (disclosed) | Series D 3rd close + 2nd $100M tender offer | BlackRock, NVIDIA, Salesforce, Deutsche Telekom and 30+ celebrity investors join; $500M ARR surpassed |
| 2026-07-02 | Bloomberg reports early tender-offer talks | Reported target: ~$22B valuation, secondary sale |
| 2026-09-24 | TechCrunch CEO interview | $22B mark confirmed as attached to the tender; ARR pacing ~$600M |
| 2026-09-28 | Eleven v4 / v4 Turbo launch | New flagship model, #1 on Artificial Analysis' Voice Arena (this review's own verification date) |
G2 loves the voices. Trustpilot is genuinely angry about billing.
3.0 out of 5, split 48% five-star vs. 38% one-star. Praise: "the voices sound remarkably natural, expressive, and realistic, with excellent pronunciation and emotional range." Complaints cluster on billing: "you spend months paying for credits and then if you unsubscribe you have your credits stolen," and on support: "their customer support is virtually non-existent -- basically zero."
A materially friendlier 4.5/5 average, with praise focused on "human-like, natural voice quality" described as unmatched in the category, and ease of use for cloning workflows -- a platform-audience gap (evaluators vs. complainers), not necessarily a different underlying product experience.
Explicitly concludes: "the gap between G2 and Trustpilot ratings is driven almost entirely by billing complaints, not product failures." Quotes a user: "I just found out that if I downgrade my plan, ElevenLabs will delete tens of thousands of credits that I already paid for." Also: "credit system is confusing and burns faster than the advertised numbers suggest."
Seven tiers, from a real free plan to custom Enterprise -- with a credit system that's the single biggest source of complaints above.
Starter
30,000 credits/mo ($5/mo billed annually).
- Commercial license unlocked
- Instant voice cloning, 20 projects
- Dubbing studio access
Creator
121,000 credits/mo ($11 first month; $18.33/mo annual).
- Everything in Starter
- Professional voice cloning unlocked
Pro
600,000 credits/mo ($82.50/mo annual).
- Everything in Creator
- 44.1kHz PCM audio via API, 192kbps quality
Scale
1.8M credits/mo, 3 seats ($249.17/mo annual).
- Everything in Pro
- Team collaboration, 3 professional voice clones
A real Free plan exists below Starter: $0/mo, 10,000 credits/mo, 3 projects, basic TTS/STT/SFX/voice design/music. Above Scale: Business is $990/mo ($825 annual), 6M credits/mo, 10 seats, 10 voice clones, and low-latency TTS as low as 5¢/minute. Enterprise is custom-priced and is the only tier that unlocks SOC 2/HIPAA/Zero Retention Mode/custom SSO (see Security). 1 credit ≈ 1 character of standard text-to-speech; other products (speech-to-text, dubbing, sound effects) spend credits at their own rates.
Your answer depends on how much voice-agent minutes and compliance actually matter to you.
Free plan -- 10,000 credits/mo is enough to evaluate the voices and cloning before paying anything.
Creator, $22/mo ($11 first month) -- the tier most individual users should start on; unlocks professional voice cloning.
Pro, $99/mo -- 44.1kHz PCM via API and 600,000 credits/mo.
Look at the Agents-specific pricing (per-minute, separate from the base plan credits) before choosing a tier -- see the Economics section below.
Enterprise, custom -- these are gated features, not available on any self-serve plan including Business.
ElevenLabs isn't the cheapest option -- see Cartesia and OpenAI in Alternatives, both materially lower-priced per character.
The thing that actually differentiates ElevenLabs right now: a new #1-ranked model that shipped mid-review.
On Sep 28, 2026 -- the day this review's research was conducted -- ElevenLabs launched Eleven v4 and v4 Turbo, its newest text-to-speech and voice-cloning models, confirmed directly on the official v4 launch page and corroborated by TechCrunch's same-day coverage. The headline claims: voice cloning from just 10 seconds of reference audio (down from longer requirements in v3), support for 90+ languages (up from 70+ in Eleven v3), better voice-identity consistency across long passages, expanded inline emotional-direction tags, and materially lower latency aimed specifically at voice agents -- v4 Turbo reports a median inference latency of roughly 100 milliseconds. Independent tracker Artificial Analysis ranked v4 #1 on its Provider Voice Arena leaderboard (Elo 1,319 from 1,674 blind-listener samples) as of this check.
| Capability | Eleven v3 (Jun 2025) | Eleven v4 / v4 Turbo (Sep 28, 2026) |
|---|---|---|
| Languages | 70+ | 90+, including Cantonese, Mongolian, Odia; notable quality gains in Japanese, Brazilian Portuguese, Mandarin, Cantonese |
| Voice cloning | Longer reference audio required | As little as 10 seconds of audio |
| Latency (Turbo) | Not optimized for real-time agents | ~100ms median inference latency |
| Independent ranking | N/A at time of launch | #1, Artificial Analysis Provider Voice Arena, Elo 1,319 |
| Launch API pricing | Standard rate ($50-$100/M chars) | 2-week launch offer: $22/M chars (v4), $11/M chars (Turbo) |
- The #1 independent-benchmark ranking is corroborated by both a specialist tracker (Artificial Analysis) and TechCrunch's own reporting, not just ElevenLabs' own marketing claim.
- 10-second voice cloning is a genuine speed improvement that materially lowers the bar for both legitimate and, worth noting, potentially abusive use.
- Shipping a new #1-ranked model the same week a $22B valuation story broke suggests real product momentum, not just a funding headline.
- The v4/v4 Turbo launch API pricing ($22 and $11 per million characters) is explicitly a 2-week introductory offer -- confirm current rates before budgeting a production workload around it.
- Faster, lower-friction voice cloning cuts both ways: it also lowers the bar for the exact kind of misuse covered in the Security section below.
Text or prompt in, produced audio out -- the review step is where quality and consent both get decided.
The human-review step is where the Trustpilot credit complaints actually originate: re-generating a line to fix pronunciation or pacing consumes credits each time, and unlike some competitors, ElevenLabs' plan credits do not appear to carry a published rollover guarantee across a downgrade -- see Pricing and the Limitations section for the specific complaint pattern.
Official ElevenLabs tutorial demonstrating exactly what the Voice Agents section below describes.
Published by the official "ElevenLabs" YouTube channel (youtube.com/@Elevenlabs), checked 2026-09-29.
The part of the business growing fastest: agents that talk on the phone, not just generate a file.
ElevenAgents is ElevenLabs' Conversational AI product: you describe or configure an agent, connect it to a phone number, chat widget, email flow, or WhatsApp, and it holds real-time voice/text conversations, calls tools, and can take action rather than just reading a script. As of the September 2026 changelog, agents support call queueing with hold audio for callers when at concurrency limits, and broader underlying-model support including gemini-3.8-flash and the music_v2_5 model. ElevenLabs also offers a dedicated HIPAA-eligible answering-service configuration for healthcare use cases, gated to Enterprise (see Security).
- Deploys across phone, chat, email, and WhatsApp from one agent definition, not four separate integrations
- Included call minutes scale with plan tier (15 min/mo on Free up to 12,375 min/mo on Business)
- Native MCP server support means an agent can call Claude, Cursor, or custom tools as part of a conversation
- Overage is billed per minute ($0.08/min, $0.16/min burst at concurrency limit) plus LLM usage billed separately based on the model chosen -- two variable costs stacked on top of the base plan
- An agent with standing phone/WhatsApp access to real customers is a materially higher-permission surface than a text-only chatbot -- review scopes and escalation paths before granting live call access
A native MCP server, a native n8n Cloud integration, and telephony via your own provider.
ElevenLabs ships an official, open-source MCP (Model Context Protocol) server, confirmed on its own GitHub and blog, explicitly built to work with Claude Desktop, Cursor, Windsurf, and OpenAI Agents as MCP clients -- meaning you can both let an ElevenLabs agent call external MCP tools, and manage your ElevenLabs agents directly from Claude or another MCP client via the hosted MCP server. Separately, ElevenLabs is a native integration inside n8n Cloud (confirmed on ElevenLabs' own blog), covering text-to-speech, voice cloning, and speech-to-text as no-code workflow nodes. Telephony for voice agents runs through a provider you connect yourself (Twilio or a SIP trunk), which bills the carrier side of the call separately from your ElevenLabs plan.
| Integration | What it does | Where it's documented |
|---|---|---|
| Official MCP server | Lets Claude/Cursor/Windsurf/OpenAI Agents call ElevenLabs tools, or manage ElevenLabs agents from an MCP client | github.com/elevenlabs/elevenlabs-mcp (official) |
| n8n Cloud | Native TTS, voice cloning, and speech-to-text nodes inside n8n automations | elevenlabs.io/blog/elevenlabs-on-n8n-cloud (official) |
| Telephony (Twilio/SIP) | Connects voice agents to real phone numbers; carrier costs billed by the provider, not ElevenLabs | elevenlabs.io/pricing/agents (official) |
Two separate cost meters -- base-plan credits, and per-minute agent minutes -- both billed on top of each other.
| What you're doing | What you pay |
|---|---|
| Evaluating on the Free plan | $0 -- 10,000 credits/mo, resets monthly |
| Generating TTS on Starter/Creator/Pro | $6-$99/mo -- 30,000 to 600,000 credits/mo included |
| Overage beyond plan credits (Flash model) | $0.05 per 1,000 characters ($50/M) |
| Overage beyond plan credits (Multilingual model) | $0.10 per 1,000 characters ($100/M) |
| Eleven v4 / v4 Turbo (2-week launch pricing, Sep 2026) | $22/M characters (v4), $11/M characters (Turbo) -- introductory, expected to revert |
| Voice-agent minutes beyond plan allowance | $0.08/min ($0.16/min burst at concurrency limit) + LLM cost billed separately |
| Transcription (Scribe) | From $0.22/hour |
| Dubbing | $0.33-$2.20/minute depending on tier |
| Rolling out to a company | Enterprise, custom -- credits, SSO/SCIM-equivalent controls, and compliance bundled into a contract |
Genuine safety infrastructure, genuine certifications on Enterprise -- and a legal history worth knowing before you clone a voice.
| Control | What's actually in place |
|---|---|
| Certifications | SOC 2 Type II, ISO 27001, PCI DSS Level 1 (per ElevenLabs' Trust Center, corroborated by independent security reviews) |
| HIPAA | Available on Agents/Enterprise only -- requires a signed BAA and Zero Retention Mode before sending PHI |
| GDPR / data residency | GDPR-compliant; EU Data Residency option available |
| Encryption | End-to-end encryption for data in transit |
| Voice-cloning consent | Recorded verbal consent statement required for Professional Voice Cloning; general/Instant cloning relies more on self-attestation (a checkbox), per Consumer Reports' 2025 study of 6 vendors |
| Enterprise gating | SOC 2/HIPAA/Zero Retention Mode/EU residency are Enterprise-tier features plus extra setup -- not available on Free through Business |
Worth separating three things here: the current safety infrastructure (Prevent/Detect/Enforce/Inform layers, the public AI Speech Classifier, C2PA participation) is real and independently checkable. The pending BIPA lawsuit is exactly that -- pending, unresolved, alleging harm ElevenLabs disputes -- and the 2024 robocall incident is a case of ElevenLabs' technology being misused by an independent third party who was separately prosecuted (and acquitted), not a case where ElevenLabs itself was found liable. All three are worth knowing before you decide how much sensitive or third-party-identity work you're comfortable putting through the platform.
Best for people who need genuinely great voice quality now, not people who need guaranteed credit rollover or self-serve compliance.
| Situation | Why ElevenLabs fits | Likely plan |
|---|---|---|
| Content creator narrating videos, podcasts, or audiobooks | Best-in-category voice quality plus ElevenReader distribution for finished audiobooks | Free, then Creator |
| Startup building a customer-facing voice agent | ElevenAgents covers phone/chat/email/WhatsApp from one definition; native MCP + n8n for the rest of the stack | Creator/Pro, then Scale for a team |
| Developer already building on Claude or via MCP clients | Official MCP server manages agents two ways -- build with ElevenLabs, or manage it from Claude | Pro or Scale |
| Healthcare org needing a compliant answering service | HIPAA-eligible Agents configuration exists, but only on Enterprise with a signed BAA | Enterprise, custom |
| High-volume TTS at the lowest possible per-character cost | Not ElevenLabs' strength -- see Cartesia/OpenAI in Alternatives | Consider a cheaper commodity API instead |
Four real gaps, not manufactured ones.
The single most consistent complaint across Trustpilot and independent reviews: users report losing tens of thousands of already-paid-for credits when they downgrade a plan, and describe credits as effectively "stolen" after unsubscribing.
SOC 2, HIPAA, Zero Retention Mode, EU Data Residency, and custom SSO all require the custom-priced Enterprise tier plus extra setup -- there's no self-serve path to any of them, even on the $990/mo Business plan.
The May 2026 Illinois BIPA class action is unresolved as of this check. It follows an earlier, similar lawsuit (Vacker v. Eleven Labs) that settled confidentially in August 2025 -- this is not a one-off allegation.
Independent of billing, several Trustpilot reviews describe support as close to unreachable ("basically zero") -- a materially different picture than G2's friendlier average.
ElevenLabs isn't the cheapest voice AI, and it isn't optimized for every tradeoff.
| If you mostly need... | Compare ElevenLabs with... |
|---|---|
| The lowest possible real-time latency for a live voice agent | Cartesia -- its Sonic 3.6 model topped a separate blind-listener latency-focused arena in Sep 2026 |
| The cheapest per-character TTS and you're already on OpenAI's agent stack | OpenAI's Realtime API / gpt-4o-mini-tts -- see our OpenAI Agents API review and OpenAI Agents SDK review |
| A turnkey voice-agent platform with a Meta-backed roadmap | PlayAI (formerly Play.ht, acquired by Meta Jul 2025) |
| An alternative cloning/generation vendor outside the big names | Resemble AI |
| A full customer-service agent platform beyond just voice | Decagon -- notably, its voice layer was originally built on top of ElevenLabs itself |
| The cheapest possible commodity TTS bundled into cloud billing you already have | Google Cloud TTS (Chirp 3 HD/Neural2/WaveNet) or Azure Neural -- both materially cheaper per character, with a smaller voice library and less cloning depth |
| Vendor | Approx. price per million characters |
|---|---|
| Google WaveNet | $4 |
| OpenAI gpt-4o-mini-tts | ~$16 |
| Google Neural2 / Azure Neural | $16 |
| Azure Neural HD | $22 |
| Google Chirp 3 HD / Murf API | $30 |
| Cartesia Scale | ~$37 |
| ElevenLabs v4 Turbo (2-week launch price) | $11 |
| ElevenLabs v4 (2-week launch price) | $22 |
| ElevenLabs standard Flash / Multilingual | $50 / $100 |
No affiliate relationship shapes this review.
JAVIS has not established a publisher affiliate program with ElevenLabs. Every CTA in this article points to the official product.
Go to the official page.
Open ElevenLabsRead the alternative that matches your real question.
Choose the next articleElevenLabs makes the most sense next to the AI agent platforms and automation tools it plugs into.
ElevenLabs is a new category for JAVIS -- AI voice -- with no direct same-category review yet. The most genuinely relevant next reads are the platforms ElevenLabs actually integrates with (Claude, via its own official MCP server; n8n, via a native Cloud integration) and the agent stack it competes with most directly (OpenAI's).
Each card image is the same real, already-verified official asset used on that target article itself (a GitHub-generated OG card for n8n; OpenAI's own Discourse-hosted announcement images for the two Agents reviews).
Questions people are actually searching right now
Is ElevenLabs owned by another company?
No. ElevenLabs is independent and still led by its 2022 co-founders, Mati Staniszewski (CEO) and Piotr Dabkowski (CTO). It has raised outside venture funding across five primary rounds plus two employee tender offers, but has not been acquired.
How much does ElevenLabs actually cost?
A real Free plan exists (10,000 credits/mo). Paid plans run $6-$990/month (Starter/Creator/Pro/Scale/Business), or lower with annual billing (2 months free), plus a custom-priced Enterprise tier. Voice agents add separate per-minute costs beyond the plan's included minutes. All figures verified live on elevenlabs.io/pricing, Sep 28-29, 2026.
What is ElevenLabs actually worth?
Its last confirmed primary funding round (Feb 4, 2026, $500M Series D led by Sequoia) valued it at $11 billion. As of Sep 24, 2026, backers reportedly mark it at $22 billion tied to an in-progress secondary tender offer where existing shareholders sell shares -- a different kind of event than a new primary raise, though widely reported as a "valuation."
What is Eleven v4?
ElevenLabs' newest flagship text-to-speech and voice-cloning model, launched Sep 28, 2026 alongside a faster "v4 Turbo" variant. It supports 90+ languages, can clone a voice from 10 seconds of audio, and was ranked #1 on Artificial Analysis' Provider Voice Arena (Elo 1,319) shortly after launch.
Is ElevenLabs safe to use for voice cloning?
ElevenLabs requires a recorded verbal consent statement for its Professional Voice Cloning tier and runs a multi-layer safety program (red-teaming, blocked celebrity-voice cloning, a public AI Speech Classifier, abuse reporting). That said, a 2025 Consumer Reports study found general/Instant cloning across most vendors, including ElevenLabs, relies more on self-attestation than robust identity verification, and ElevenLabs is currently a defendant in a pending May 2026 federal lawsuit over training-data consent -- worth reading before you clone a voice that isn't clearly your own.
Was ElevenLabs involved in the 2024 Biden robocalls?
Its voice-cloning technology was used by a third-party political consultant, Steve Kramer, to create the calls. Kramer was acquitted on all criminal charges by a New Hampshire jury; the telecom carrier that transmitted the calls separately paid a $1M FCC settlement. ElevenLabs itself was not a party to either case or fined.
Does ElevenLabs work with Claude or MCP?
Yes. ElevenLabs ships an official, open-source MCP server confirmed to work with Claude Desktop, Cursor, Windsurf, and OpenAI Agents -- letting an ElevenLabs agent call external MCP tools, or letting you manage ElevenLabs agents directly from an MCP client like Claude.
How does ElevenLabs compare to Cartesia and OpenAI's voice tools?
Independent 2026 comparisons generally rank ElevenLabs highest on overall voice quality and the broadest voice marketplace/cloning depth, Cartesia highest on real-time streaming latency for live agents, and OpenAI as the cheapest mainstream option with strong instructable voice character, especially for teams already building on OpenAI's Agents/Realtime stack.
Where this came from, and when it was checked.
Pricing (Free/Starter/Creator/Pro/Scale/Business/Enterprise): elevenlabs.io/pricing and elevenlabs.io/pricing/agents, Sep 28-29, 2026; cross-checked against 5 independent trackers (HappyRobot, Flexprice, WebsiteVoice, layer3labs, Cekura) -- all agree on exact figures.
Product homepage claims, safety architecture: elevenlabs.io and elevenlabs.io/safety, official, checked Sep 28-29, 2026.
Company background, founders, HQ, funding timeline through Series C: en.wikipedia.org/wiki/ElevenLabs, cross-checked against ElevenLabs' own official blog posts for the Series D and tender-offer entries.
Series D ($500M, $11B valuation, Feb 4 2026): elevenlabs.io/blog/series-d official announcement; corroborated by TechCrunch and CNBC same-day coverage.
$500M ARR + third Series D close + second tender offer: elevenlabs.io/blog/500m-arr-and-new-investors, official, published May 5 2026, updated Sep 20 2026.
$22B reported valuation: Bloomberg (Jul 2, 2026, early tender-offer talks) and TechCrunch (Sep 24, 2026, CEO interview confirming the mark); both explicitly framed as a secondary-tender valuation, not a new primary raise.
Eleven v4 / v4 Turbo launch (Sep 28, 2026): elevenlabs.io/v4 official page; TechCrunch same-day coverage; runtimewire.com and Artificial Analysis for the #1 Provider Voice Arena ranking claim.
MCP server: github.com/elevenlabs/elevenlabs-mcp (official repo) and elevenlabs.io/blog/introducing-elevenlabs-mcp (official announcement).
n8n integration: elevenlabs.io/blog/elevenlabs-on-n8n-cloud and elevenlabs.io/agents/integrations/n8n, both official.
Compliance/security (SOC 2, HIPAA, GDPR, Zero Retention): compliance.elevenlabs.io Trust Center, corroborated by independent reviews from Deepgram and Nudge Security.
Voice-cloning consent practices, industry safety gap: BankInfoSecurity coverage of a 2025 Consumer Reports study of 6 voice-cloning vendors including ElevenLabs.
Pending BIPA lawsuit (filed ~May 2026): Sifted (sifted.eu/articles/elevenlabs-lawsuit-2026), independent reporting; status: pending as of this check.
Vacker v. Eleven Labs settlement (Aug 2025): p4sc4l.substack.com independent legal-analysis write-up.
2024 NH Biden robocall incident: CBS News/NBC Boston reporting on Steve Kramer's criminal acquittal; NBC News reporting on Lingo Telecom's separate $1M FCC settlement; Wikipedia's ElevenLabs controversies section for the original technology-attribution claim.
Trustpilot rating (3.0/5, 1,165 reviews): direct read of trustpilot.com/review/elevenlabs.io, Sep 29, 2026.
G2 rating (~4.5/5, 1,140 reviews): search-aggregated read across multiple G2 subpages; JAVIS could not read g2.com directly.
G2-vs-Trustpilot gap analysis, specific complaint quotes: aiprotivity.com/elevenlabs-review/, independent review, published Jun 19 2026, updated Jul 2026.
Vendor per-character price comparison: synthesized from multiple independent pricing-tracker sources (~mid-2026 rate snapshot), cross-checked against ElevenLabs' own documented overage rates.
ElevenReader (200,000+ titles, $11/mo): Tech Times, published May 22, 2026.
Official video: youtube.com/watch?v=fnivYSh0Cqk, authorship confirmed on YouTube (author "ElevenLabs", channel @Elevenlabs).
