Pella Research

6.4easy start · #15 of 20
in AI Deepfake Detection Tools

Free to practise onnot on record
Free trialnot on record
Well documentedyes
Runs where you worknot on record

Runs on api, Web.

Visit site Alternatives

Pella Research is ranked #15 of 20 in AI deepfake detection tools on The Geeks Club. It runs on API, Web.

Compared on AI deepfake detection tools

Free plan: Yes
Media types: audio
API access: Yes
Face analysis: No

Best Pella Research alternatives

See all 19

Netarx 0 of 4 in the starter kit Lesson 17.4 Deepfake Detector 0 of 4 in the starter kit Lesson 27.1 ForgeSpy 0 of 4 in the starter kit Lesson 37.1 Hive AI-Generated Content Detection 0 of 4 in the starter kit Lesson 47.1 Reality Defender 0 of 4 in the starter kit Lesson 57.1 DuckDuckGoose DeepDetector 0 of 4 in the starter kit Lesson 67.0

The record

Product: Pella Research builds models to detect synthetic speech by analyzing vocoder artifacts, spectral statistics, and prosodic cues.
Research and benchmarks: It publishes evaluation sets that track new voice cloning systems, with documented provenance and no train/test leakage.
API: The detection API is a REST endpoint that returns calibrated authenticity scores and per-segment analysis, with access granted on request.
Live demo: The browser demo lets users upload or record a voice clip and receive a verdict without signup; the page says no audio is stored.
API audience: The API is offered to journalists, platforms, researchers, and institutions, and the page lists a free tier for accredited journalists and academics.
Access response: Pella says it reviews API requests and typically responds within 24 hours.
Benchmark result: Pella reports that pellav2 achieved 95.82% accuracy, a 0.959 F1 score, and 57 ms latency in the August 2026 Podonos benchmark.
Benchmark limitation: Pella reports 93.6% accuracy on m4a files and says its errors lean toward false positives.
Real-world conditions: Pella says its models are trained and tested with phone-line compression, background noise, and re-encoding.
Model availability: The site describes pellav2 as open weights under the MIT license and says the public checkpoint reproduces the submitted benchmark score.
Demo limitation: The live demo runs on free shared hardware, and its first request after a quiet period can take up to a minute to wake up.
Intended users: Pella describes its work as serving platforms, journalists, courts, and researchers investigating whether audio is synthetic.
Location: The site lists Amsterdam, Netherlands.
Purpose: Pella Research studies synthetic speech and builds models to detect it in real-world audio.

Where it ranks on The Geeks Club

Best AI Deepfake Detection Tools in 2026#15 of 20

Is Pella Research yours?

Claim it for free: prove the domain, then correct facts, plans and screenshots. An editor reviews every change.

Claim it free Promote Pella Research

Sources

pellaresearch.com· checked 7 Oct 2026
pellaresearch.com/mission.html· checked 7 Oct 2026