AI Aware
The AI detector for text, images, video and audio
About AI Aware
AI Aware is an AI content and deepfake detector that checks four kinds of media from one platform: text, images, video and audio. You paste text or upload a document, image, video or audio file, and it returns a confidence score with the flagged passages or media highlighted. The problem it addresses is that AI-generated and manipulated content rarely arrives in a single form. The company gives the example of a fraudster sending a cloned voice note, a deepfake video and an AI-written message as part of the same attack, and most detectors only check text.
Rather than a single model, AI Aware runs an ensemble of detection models, and the company says it has tested and trained on hybrid content (AI text edited by people) and manipulated content (human work altered by AI), including text passed through "humaniser" tools. Text results are given at paragraph level with calibrated probability scores, so a document can be read section by section instead of as one overall percentage. The site reports over 97% accuracy across the four media types and 15 million+ detection tests run; these are the vendor's own figures. AI Aware is the trading name of Happy AI Ltd, a London company, which says its work is research-led with funding from Innovate UK and that it is a member of the Content Authenticity Initiative. You can run 3 free checks without signing up, and the text detector currently states it is for English only.
First you choose Text or Media in the detector. For text you paste content or upload a TXT, DOCX or PDF file; the tool asks for at least five sentences and notes it is not reliable on maths, tables or code. For media you upload a JPG, PNG, WEBP, GIF or BMP image, an MP4, MOV, WEBM, AVI or MKV video, or an MP3, WAV, OGG or M4A audio file, or paste a direct URL to one.
Second AI Aware analyses the content with multiple detection models. Text is checked paragraph by paragraph, video is checked frame by frame, and audio is examined with spectral analysis for cloned voices.
Third you read the results. For text you see an overall verdict, the percentage of the document containing AI, a confidence level and a segment chart showing the probability of AI through the document, with AI-written passages highlighted. Image and video results show an AI-generated score and a frame-by-frame deepfake score.
Each check uses credits once the free checks are spent: 3 for a text check up to 5,000 words, 1 for an image, 5 for audio and 10 for video.
- •AI Text Detector - Detects text from ChatGPT, Claude, Gemini, Grok, LLaMA and other models in essays, reports, CVs and legal documents, separating AI, human and hybrid writing at paragraph level.
- •AI Image Detector - Checks photos and artwork for output from generators such as Midjourney, DALL-E, Stable Diffusion and FLUX, looking at pixels, texture, lighting and metadata.
- •Deepfake Detector - Scores video for deepfakes, face swaps and AI-generated footage frame by frame, trained on output from generators including Synthesia, Sora, Runway, HeyGen and Seedance.
- •AI Audio Detector - Flags cloned voices and AI-generated speech, covering voice cloning tools such as ElevenLabs.
- •Visualisation Tool - Shows how much of a document is AI-generated and whether it sits in blocks or is spread through the text, so a result can be judged in context.
- •Team and API Access - Volume pricing, API access and tailored integrations are offered to organisations through the sales team.
AI Aware is aimed at organisations that need to know whether content is real. Its site lists cybersecurity, financial services, legal, recruitment, publishing and education. Fraud and security teams can check suspicious calls, voice notes and videos for cloning and deepfakes. Legal and compliance teams can screen contracts, witness statements, funding applications and regulatory submissions, where formulaic legal language can trip simpler detectors.
Recruiters can check CVs and application text in volume, and educators can review essays and reports with paragraph-level results rather than a single score. Journalists and publishers can verify images, footage and audio before running them.
Individuals with an occasional check can use the free checks and buy one-off credit packs that do not expire, while regular users can take the monthly plan and firms can ask for volume pricing.
Pricing
Priced in another currency
- StarterNot listed
- PlusNot listed
- ProNot listed
- Monthly9.99 GBP/mo
- TeamContact sales
From the vendor pricing page, 2026-11-10
