Pindrop Pulse vs Resemble Detect: voice deepfake detection on the same 14 fields and the voice fields
Resemble Detect publishes a price and Pindrop Pulse does not: Resemble's Flex plan has no subscription fee and bills $0.035 a second of audio, with a Team plan at $350 a month that cuts the rate to $0.015 a second, while every Pindrop Pulse page ends in "Schedule a demo" or "Talk to a human" and states no price for any of the three Pulse products, by the vendor pages read 9 November 2026 (resemble.ai/pricing, pindrop.com/pulse-for-contact-centers). Neither card has a study record: none of the 35 records in the studies index scored either product, and the two speech challenges in the index, ASVspoof 5 and SAFE 2025, score anonymised teams, with Pindrop appearing in ASVspoof 5 as an organiser affiliation and workshop co-sponsor rather than as a participant (arXiv 2601.03944, author affiliations and footnote 8, read 9 November 2026). The two cards are read side by side below on the same 14 fields and then on the fields a voice fraud or trust-and-safety buyer asks about first: modalities, input limits and formats, retention, deployment, compliance, whether a price is published, and whose number each accuracy claim is.
Updated Fields checked 9 November 2026
Short answer
Both score voice and video and neither scores text; Resemble Detect adds images and music, 19 upload extensions with a 150 MB direct-upload ceiling, 54 stated languages and a per-second price, while Pindrop Pulse states no file format, size or duration limit, 40 or more languages with none named, and no price (docs.resemble.ai, resemble.ai/benchmarks, pindrop.com/company-news/pindrop-launches-pulse-inspect, read 9 November 2026). Pindrop's privacy notice keeps derived "voice features" for up to three years and its Meetings DPA destroys personal data within 90 days of a contract's end; Resemble's policy deletes submitted media after analysis unless it is flagged and offers a zero retention mode (pindrop.com/legal/privacy-notice-for-pindrop-solutions, resemble.ai/privacy-policy, read 9 November 2026). Every accuracy figure on this page is the vendor's own: Pindrop's 99% (accuracy) against known engines is footnoted to "the ASVspoof5 dataset of 32 engines" and Resemble's 99.5% (accuracy) to the Podonos Audio DFD Bench, and no indexed study scored either. No detector score should be the only evidence in a misconduct, hiring or publishing decision.
How do Pindrop Pulse and Resemble Detect compare on the 14 fields?
| Field | Pindrop Pulse | Resemble Detect |
|---|---|---|
| 1. Name, vendor, official domain | Pindrop Pulse, Pindrop Security, Inc., pindrop.com | Resemble AI Detect, Resemble AI, Inc., resemble.ai |
| 2. Modalities detected | voice (audio); video (Pulse for Meetings: "fake video and audio, synthetic avatars, and face swaps") | voice (audio), image, video, music (Suno and Udio listed as covered) |
| 3. Generators covered | No named list: the model is "trained on 350+ deepfake generation tools, 20M+ unique utterances and 40+ languages" (Pulse Inspect release); the Meetings FAQ cites "a proprietary dataset of 20 million audio samples representing more than 120 AI voice generation systems"; "known" engines are distinguished from "new or unseen" ones | Audio: ElevenLabs, OpenAI TTS, Azure TTS, Google TTS, AWS Polly, Suno, Udio, PlayHT, Murf, Descript Overdub, Bark, VALL-E, VALL-E 2, YourTTS, Tortoise TTS, XTTS, StyleTTS 2, MetaVoice, Kokoro, VoiceBox, NaturalSpeech 3, HierSpeech++ "and over 100 more"; image and video lists from DALL-E 3 to HunyuanVideo; "250+ Generative AI models" |
| 4. Languages supported | 40 or more stated ("40+ languages, covering over 90% of the internet's spoken languages"), none named | 54 stated ("54 languages detected from Arabic to Spanish"), 30 named, "Validated against MLAADv10" |
| 5. Input limits | Pulse for Contact Centers scores live calls and "can detect deepfake audio in just two seconds" of speech; Pulse Inspect takes uploaded audio and returns continuous scores "every four seconds" through a batch API with adjustable thresholds; Pulse for Meetings "continuously analyzes meeting audio and video"; no file size, duration or format limit published | API: direct multipart upload up to 150 MB, a public HTTPS URL with no size limit, a token for larger files; 19 extensions from .wav and .mp3 to .mkv and .webp; audio and video scored in windows of 1 to 4 seconds; batch and WebSocket streaming endpoints; a zero_retention_mode flag |
| 6. Free tier | None published; the Pulse Meeting Terms define a Free Trial Product "as reflected in the relevant Order", for internal evaluation only | Flex at $0 a month with no subscription fee, prepaid credits that never expire; "Get your first detection for free" with no stated allowance |
| 7. Paid plans | Pulse for Contact Centers, an add-on to Pindrop Protect or Passport, with a "$1M deepfake detection warranty" for eligible 3-year subscriptions; Pulse for Meetings through a marketplace, reseller or direct order, fees in US dollars due within 30 days; Pulse Inspect in Preview since 15 August 2024; no price published for any | Flex $0 a month: audio $0.035 a second, images $0.035, video $0.07 a second; Team $350 a month or $280 a month billed annually: audio $0.015 a second, 5 seats; Business $1,000 a month or $800 billed annually: 20 seats, SSO; Enterprise by contact |
| 8. API | Yes: Pulse Inspect is "an API-driven batch-processing platform"; no public API documentation page was found on 9 November 2026 | Yes, docs.resemble.ai/detect: REST with Python, Node.js and JavaScript SDKs |
| 9. Integrations | Contact-centre platforms Verizon, Genesys, Five9, Amazon Connect, Cisco and NICE; Zoom (including Zoom Contact Center), Cisco Webex and Microsoft Teams for meetings; a hiring platform integration that writes results back to the candidate record; Pindrop Protect and Passport | REST API and SDKs, batch and WebSocket endpoints; MCP for Cursor and Claude Code; a Chrome extension; Microsoft Teams, Zoom, Google Meet, Webex and phone systems for real-time meeting detection; on-premise, air-gapped and private VPC deployment (AWS, GCP, Azure) with Docker and Kubernetes; Carahsoft contract vehicles |
| 10. Vendor accuracy and false positive claims | "99% Detection accuracy against known engines", footnoted "Pindrop Pulse accuracy computed on the ASVspoof5 dataset of 32 engines"; "<1% false positive rate", footnoted "Pindrop Product Data, 2024"; "up to 99.4%" (accuracy) with the authentication platform; 99% on known and over 90% on unseen engines on the 20 million sample dataset; 81 of 84 samples in an NPR test (96.4%); a Podonos ranking of June 2026 | DETECT-World "Up to 99.5%" (accuracy); benchmarks page (August 2026): 99.5% pooled detection accuracy on the Podonos Audio DFD Bench with "<1% False positive and false negative rate"; audio false negative rate 0.4% on 4,524 files; images 95.8% (accuracy) with a 4% false positive rate on 1,000 files; video 98.2% (accuracy) with 0.95% on 1,000 files; "#1 Ranked on third-party benchmarks" |
| 11. Independent studies | 0 records; a note on the card records that Pindrop is an ASVspoof 5 organiser affiliation and co-sponsor and that the challenge papers print no Pindrop result | 0 records |
| 12. Data retention and privacy statement | Privacy Notice (27 March 2026): content-agnostic "voice features" extracted from customer recordings "are retained by Pindrop for a period of up to three (3) years"; "Customer voiceprints are not retained"; Pulse for Meetings DPA: after a 30-day return period, personal data destroyed "no later than 90 days"; Support Data may be retained for product optimisation and that right survives termination | Privacy Policy (14 August 2026), section 11: "Audio, image and video submitted for detection analysis is deleted after analysis unless it is flagged"; flagged participant recordings kept as the business customer directs and no longer than three years; deletion requests completed within 30 days; Zero Retention Mode deletes media after analysis |
| 13. Launch year and ownership | 2024 (Pulse Inspect launched 15 August 2024); Pindrop Security, Inc., Atlanta, "a global leader in voice authentication and security since 2011"; terms governed by Delaware law; no parent stated | Not stated; Resemble AI, Inc., Mountain View, California, a company that "produces synthetic voices and provides synthetic-media detection services"; hosted in the United States on Google Cloud, Render and Cerebrium; no parent stated |
| 14. Rankings the card appears in | Studies by content type (0 records), languages (40), launch year, platform integrations | Studies by content type (0 records), languages (54), file types (19 formats), generators covered, platform integrations |
Which studies tested Pindrop Pulse or Resemble Detect?
None. No record in the studies index ran either product, and the index enters no vendor figure as a study result. ASVspoof 5 scored 53 anonymised teams on 138,688 bona fide and 542,086 spoofed utterances and found the best open-condition detector at a 2.59% equal error rate; Pindrop is in the paper as co-author Hemlata Tak's affiliation and as co-sponsor of the 2024 workshop, not as a scored team, and the evaluation paper prints no Pindrop result (arXiv 2601.03944, author list and footnote 8; arXiv 2408.08739, Table 4 and Acknowledgements, read 9 November 2026). SAFE 2025 scored twelve teams on 2,600 clips from 13 text-to-speech models and names neither vendor as participant, baseline or sponsor (arXiv 2510.03387, Table 5 and Acknowledgments, read 9 November 2026). The two vendors cite each other's benchmark: Resemble's benchmarks page lists Pindrop at 95% among the Podonos figures beside its own 99.5%, and Pindrop's page cites a Podonos study of 17 systems from June 2026 that "ranked Pindrop as the fastest most accurate commercial detector evaluated" (resemble.ai/benchmarks, pindrop.com/pulse-for-contact-centers, read 9 November 2026); the index has not read the Podonos report and carries neither figure as evidence. Both cards therefore sit at 0 records in studies by content type, and the voice fields below are the comparison.
| Voice field | Pindrop Pulse | Resemble Detect |
|---|---|---|
| Modalities | Voice (audio) on live calls and uploaded files; video inside meetings | Voice (audio), image, video, music |
| Input limits and formats | No file size, duration or format limit published; live calls scored within two seconds of speech; uploads scored every four seconds | 19 extensions: wav, mp3, m4a, ogg, aac, flac, amr, 3gp, 3gpp, mp4, mov, avi, mkv, webm, jpg, jpeg, png, gif, webp; 150 MB direct upload, URL with no limit, 200 MB with watermark detection; 1 to 4 second windows |
| Retention of submitted audio | Derived voice features kept up to three years; voiceprints not retained; Meetings data destroyed within 90 days of a contract's end after a 30-day return period; no period stated for the recordings themselves | Deleted after analysis unless flagged; flagged recordings kept as the customer directs and at most three years; zero_retention_mode deletes the file after detection and redacts media URLs |
| Deployment | Cloud on Verizon, Genesys, Five9, Amazon Connect, Cisco and NICE platforms; inside Teams, Zoom and Webex; on-premises customers named in the privacy notice | Cloud API and web app; on-premise, air-gapped and private VPC on AWS, GCP or Azure; Docker and Kubernetes; Carahsoft for US public-sector contracts |
| Compliance standards named | None named on the pages read | SOC Type 2 documentation on the Enterprise plan |
| Pricing published | No: demo and sales forms for all three Pulse products; a $1M warranty on eligible 3-year subscriptions | Yes: Flex $0 a month at $0.035 a second of audio; Team $350 a month at $0.015 a second; Business $1,000 a month; Enterprise by contact |
| Vendor claim and whose number it is | 99% (accuracy) against known engines, Pindrop's computation on the ASVspoof5 dataset; under 1% false positives, "Pindrop Product Data, 2024"; 96.4% in an NPR test of 84 samples, as Pindrop reports it | Up to 99.5% (accuracy), Podonos Audio DFD Bench as Resemble reports it; 0.4% audio false negative rate on 4,524 files, Resemble's test sets; EER figures from Hugging Face arena leaderboards of March 2026, as Resemble reports them |
| Verdict format | "A probabilistic determination as to whether a Meeting Participant may be human or a machine"; continuous scores every four seconds on uploads; adjustable thresholds | A score per window with the best-guess generator first; "There is no confidence cut-off: the first entry is always the best guess, and its prob shows how sure the model is" |
Fields as recorded on the two cards from the vendor pages named above, read 9 November 2026; the data retention ranking ranks neither card, because neither clause states a period for the submitted recording itself, the compliance view records Resemble's SOC Type 2 line and no standard for Pindrop, and the languages ranking carries Resemble's 54 and Pindrop's 40.
What do the two cost over 12 months?
| Plan | Monthly as stated | Annual as stated | Inclusions as stated | Checked on |
|---|---|---|---|---|
| Pindrop Pulse for Contact Centers | No public price | No public price | Add-on to Pindrop Protect or Passport; synthetic voice detection on live calls in the IVR and with agents; "$1M deepfake detection warranty" for eligible 3-year subscriptions to the full suite | 9 November 2026 |
| Pindrop Pulse for Meetings | No public price | No public price | Audio and video deepfake, voice match and location checks inside Teams, Zoom and Webex; bought through a marketplace, reseller or direct order; Free Trial Product possible per order | 9 November 2026 |
| Pindrop Pulse Inspect | No public price | No public price | File upload and batch API with scores every four seconds; in Preview since 15 August 2024 for media, nonprofits, governments and social networks | 9 November 2026 |
| Resemble Detect Flex | $0, no subscription fee | Prepaid credits that never expire: audio $0.035 a second, images $0.035 each, video $0.07 a second; 1 seat; platform and API access; real-time meeting detection; SynthID check and C2PA signing | 9 November 2026 | |
| Resemble Detect Team | $350 billed monthly; $280 a month billed annually ("save $840/yr") | $3,360 at the stated annual monthly rate times twelve | Audio $0.015 a second, images $0.015, video $0.03 a second; large file uploads over 150 MB; batch uploads; 5 seats | 9 November 2026 |
| Resemble Detect Business | $1,000 billed monthly; $800 a month billed annually ("save $2,400/yr") | $9,600 at the stated annual monthly rate times twelve | Team rates; org-wide calendar integration and security notifications for meetings; SSO; 20 seats | 9 November 2026 |
| Resemble Detect Enterprise | Custom | Custom | Volume pricing, SLAs, custom model training, on-premises deployment, dedicated support, SOC Type 2 documentation | 9 November 2026 |
Prices in USD as shown on resemble.ai/pricing, checked 9 November 2026; Resemble's annual figures are the index's multiplication of the stated annual monthly rate by twelve, because the page prints the saving rather than a yearly total. At Flex rates, one hour of audio costs $126 and at Team rates $54 before the subscription, by the index's multiplication of the per-second price. Pindrop publishes no price on any page read (pindrop.com/pulse-for-contact-centers, pindrop.com/legal/pulse-meeting-terms, read 9 November 2026), and its Meeting Terms state that direct-order fees are invoiced in US dollars and due within 30 days.
One of the comparisons in the compare hub. The cards are Pindrop Pulse and Resemble Detect, and the home board answers which AI detectors exist and who has tested them? A contact-centre, hiring or security team weighing a detector sold as an add-on to a voice authentication platform against one sold by the second through an API is the reader this pair is written for; the voice and video question lists every card that states both media, and the business view reads the same fields for a fraud team.
Which card fits which reader?
- Choose Resemble Detect ifyou want a published price and a self-serve start: Flex at $0 a month and $0.035 a second of audio, with an API key on sign-up (resemble.ai/pricing, read 9 November 2026). Pindrop sells every Pulse product by demo.
- Choose Pindrop Pulse ifthe threat is a live call in Genesys, Five9, Amazon Connect, NICE, Cisco or Verizon telephony: those six platforms are named and the detection runs in the IVR and with the agent within two seconds of speech (pindrop.com/pulse-for-contact-centers, read 9 November 2026). Resemble names phone systems without a platform list.
- Choose Resemble Detect ifthe audio cannot leave the building: on-premise, air-gapped and private VPC deployment on AWS, GCP or Azure are stated (resemble.ai/products/detect, read 9 November 2026). Pindrop's privacy notice names on-premises customers and its product pages state no deployment option.
- Choose Resemble Detect ifyou need the format and size limits in writing: 19 extensions, a 150 MB direct-upload ceiling and 1 to 4 second windows are documented (docs.resemble.ai/detect/create, read 9 November 2026). Pindrop publishes no file limit.
- Choose Pindrop Pulse ifthe use is a hiring interview: the hiring platform integration writes results back to the candidate record (pindrop.com/pulse-for-meetings, read 9 November 2026). Resemble covers the same three meeting tools and Google Meet with no hiring integration stated.
- Choose Resemble Detect ifthe same account must also check images, video files and music: four modalities against Pindrop's voice and in-meeting video (resemble.ai/products/detect, read 9 November 2026).
- Choose Resemble Detect ifthe submitted media must not be kept: deletion after analysis unless flagged, and a zero_retention_mode flag, against Pindrop's three-year retention of derived voice features (resemble.ai/privacy-policy, pindrop.com/legal/privacy-notice-for-pindrop-solutions, read 9 November 2026).
- IfChoose neither on accuracy: no indexed study scored either product, and each vendor's figure rests on a set the vendor chose, Pindrop's on the ASVspoof5 dataset by its own computation and Resemble's on the Podonos bench, neither of which the index has read as a result for these products.
How does the evidence compare, tool by tool?
Pindrop Pulse
- Register ID
- AIDI-D-0041
- Study records
- 0 in the index
- Shared with Resemble Detect
- 0
- Vendor accuracy claim
- Audio deepfake detection page: '99% Detection accuracy against known engines', footnoted 'Pindrop Pulse accuracy computed on the ASVspoof5 dataset of 32 engines'. Pulse for Contact Centers page: '99% (accuracy) against known engines' and '<1% false positive rate', both footnoted 'Pindrop Product Data, 2024'; FAQ: 'When combined with Pindrop's multifactor authentication platform, Pulse has been shown to deliver up to 99.4% (accuracy) with less than 1% false positives'. Pulse for Meetings FAQ: 'a 99% detection rate for known deepfake engines and over 90% detection for new or unseen deepfake generation engines, while maintaining a false positive rate of less than 1%, based on testing against a proprietary dataset of 20 million audio samples representing more than 120 AI voice generation systems'; 'In a separate, independent test by NPR ... Pulse correctly identified 81 of 84 audio samples, which Pindrop calculated as 96.4% accuracy'; the Podonos Audio Deepfake Detection Benchmark 'ranked Pindrop as the fastest most accurate commercial detector evaluated in its benchmark' (footnote: Podonos's study of 17 audio deepfake detection systems, June 2026). Pulse Inspect release: 'identifying AI-generated audio with 99% (accuracy)'. None of these pages is an index study record.
- Free tier
- None published. Every product page ends in 'Schedule a demo' or 'Talk to a human'; the Pulse Meeting Terms define a Free Trial Product made available 'as reflected in the relevant Order', 'solely for internal evaluation and testing purposes, where Meeting Participants shall be limited to Company's own internal personnel'
- Languages
- 40 stated
Resemble Detect
- Register ID
- AIDI-D-0042
- Study records
- 0 in the index
- Shared with Pindrop Pulse
- 0
- Vendor accuracy claim
- Product page: DETECT-World 'Up to 99.5% (accuracy)', 'Zero day coverage', '250+ models'; '#1 Ranked on third-party benchmarks for audio deepfake detection'. Benchmarks page (last updated August 2026): '99.5% Pooled detection accuracy, Podonos Audio DFD Bench' with '<1% False positive and false negative rate'; audio false negative rate 0.4% on 'Test sets include mp3, wav, flac, ogg, m4a, webm across 4,524 files'; images '95.8% Average accuracy ... when benchmarked against 1,000 image files'; video '98.2% Average accuracy ... when benchmarked against 1,000 video files'; the same page lists competitors' Podonos figures (Whispeak 97.7%, Aurigin AI 96.8%, Pella Research 95.8%, Pindrop 95%). Per-generator image figures on the product page run from 94% (Stable Diffusion SDXL and SD 3.5) to over 99% (StyleGAN). The Podonos benchmark is not an index study record.
- Free tier
- Flex plan at $0 a month: 'no subscription fee', credits bought as you go and never expiring, with deepfake detection for audio, image and video, Intelligence insights, meetings detection and API access on 1 seat; the product page offers 'Try it free' and 'Get your first detection for free' with no stated allowance
- Languages
- 54 stated
How this comparison is built
- Cards compared
- Pindrop Pulse and Resemble Detect
- Fields
- The 14 fields of every card, in card order
- Shared study records
- 0
- Fields checked
- 9 November 2026
- Records in the studies index
- 35 records
Both columns are drawn from the same 14 fields in the index data file, read on 9 November 2026 from the vendor page recorded on each card, and from the studies index, which holds no record for either product; the second table repeats the eight fields a voice buyer asks about first, each copied from the card field that carries it, with a row that names whose number each claim is. Where a vendor publishes no figure, the cell says so rather than carrying a number from a third party. The ASVspoof 5 record is linked because Pindrop's claim footnote names the dataset and because Pindrop is an organiser affiliation in the paper, which this page repeats without treating it as a result. The rules, the recheck schedule (next pass 8 December 2026) and the correction route are on the methodology page; the music detectors question reads Resemble's music line beside the four other cards that state it, and the modalities view lists every card by every medium it states.
Where each figure comes from
Values as each vendor states them on pindrop.com/pulse-for-contact-centers, pindrop.com/pulse-for-meetings, the Pulse Inspect launch release, the Pulse Meeting Terms and the Privacy Notice for Pindrop Solutions, and on resemble.ai/products/detect, resemble.ai/pricing, resemble.ai/benchmarks, docs.resemble.ai/detect/create and resemble.ai/privacy-policy, all read 9 November 2026. Pindrop's "32 engines" footnote and Resemble's Podonos figure are both vendor attributions: the ASVspoof 5 database holds 32 attack identifiers, A01 to A32, and the challenge papers score no Pindrop system, and the Podonos benchmark is not an index record.
Questions about Pindrop Pulse and Resemble Detect
Has any study tested Pindrop Pulse or Resemble Detect?
No indexed study has scored either product; both cards sit at 0 records in studies by content type. ASVspoof 5 and SAFE 2025 score anonymised teams, and Pindrop appears in ASVspoof 5 as an organiser affiliation and workshop co-sponsor, not as a participant (arXiv 2601.03944, read 9 November 2026).
Which publishes a price?
Resemble Detect: Flex at $0 a month with audio at $0.035 a second, Team at $350 a month, Business at $1,000 a month and Enterprise by contact (resemble.ai/pricing, read 9 November 2026). Pindrop publishes no price for Pulse for Contact Centers, Pulse for Meetings or Pulse Inspect.
What is Pindrop's "ASVspoof5 dataset of 32 engines" claim?
A vendor computation: Pindrop's audio deepfake detection page footnotes its 99% (accuracy) as "computed on the ASVspoof5 dataset of 32 engines", and the ASVspoof 5 database holds 32 attack identifiers, A01 to A32, but the challenge papers score 53 anonymised teams and print no Pindrop result (pindrop.com/pulse-for-contact-centers; arXiv 2601.03944, read 9 November 2026).
Which runs on-premise?
Resemble Detect states on-premise, air-gapped and private VPC deployment on AWS, GCP or Azure with Docker and Kubernetes (resemble.ai/products/detect, read 9 November 2026). Pindrop's privacy notice refers to on-premises customers who retain voice features "for varying periods of time", and its product pages state cloud contact-centre and meeting deployments (pindrop.com/legal/privacy-notice-for-pindrop-solutions, read 9 November 2026).
How long does each keep the audio?
Pindrop keeps derived voice features for up to three years and does not retain voiceprints, with Meetings data destroyed within 90 days of a contract's end; Resemble deletes submitted media after analysis unless it is flagged, keeps flagged recordings at most three years, and completes deletion requests within 30 days (pindrop.com/legal/privacy-notice-for-pindrop-solutions, resemble.ai/privacy-policy, read 9 November 2026). Neither states a period for the recording itself, so the data retention ranking ranks neither.
Which names more languages?
Resemble Detect states 54 and names 30; Pindrop states "40+" and names none (resemble.ai/benchmarks, pindrop.com/company-news/pindrop-launches-pulse-inspect, read 9 November 2026). No study in the index tested either in any language.
Does either name a security standard?
Resemble lists "SOC type 2 documentation" among its Enterprise inclusions (resemble.ai/pricing, read 9 November 2026); Pindrop names no standard on the pages read, so only Resemble appears in the compliance view.
Which accepts more file types?
Resemble Detect, 19 extensions on its API against no published list for Pindrop (docs.resemble.ai/detect/create, read 9 November 2026); the file types ranking counts all 19 for Resemble Detect from 10 November 2026 (17 until the amr and 3gp tokens were counted under the stated rule) and does not rank Pindrop.
Sources and revisions
Sources
- Pindrop Pulse for Contact Centers (claims, plans, integrations)pindrop.com/pulse-for-contact-centers
- Pindrop Pulse for Meetings (modalities, FAQ)pindrop.com/pulse-for-meetings
- Pindrop Pulse Inspect launch release (generators, languages, input, API)pindrop.com/company-news/pindrop-launches...
- Pulse Meeting Terms (free trial, ownership)pindrop.com/legal/pulse-meeting-terms
- Privacy Notice for Pindrop Solutions (retention)pindrop.com/legal/privacy-notice-for-pind...
- Resemble Detect product pageresemble.ai/products/detect
- Resemble AI pricingresemble.ai/pricing
- Resemble AI benchmarksresemble.ai/benchmarks
- Resemble Detect API, create a detection (formats and limits)docs.resemble.ai/detect/create
- Resemble AI privacy policyresemble.ai/privacy-policy
- Resemble AI terms of service (ownership)resemble.ai/terms-of-service
- ASVspoof 5 evaluation (Wang et al. 2026)arxiv.org/abs/2601.03944
- ASVspoof 5 challenge summary (Wang et al. 2024)arxiv.org/abs/2408.08739
- SAFE challenge (Trapeznikov et al. 2025)arxiv.org/pdf/2510.03387
