Artistic analyzes speech recordings and conversation transcripts to surface early markers of autism spectrum disorder — with transparent, explainable AI built for clinicians and researchers.
How it works
Drop an audio recording (WAV, MP3, FLAC) or a CHAT transcript file from your session directly into the tool — no conversion or formatting needed.
Our AI examines conversation flow, voice patterns, and language structure — then combines all signals into a single, calibrated ASD screening result.
See an annotated transcript, a plain-language explanation of what drove the prediction, and explore what-if scenarios to understand the result in depth.
What we analyze
How your child takes turns, stays on topic, and responds in back-and-forth dialogue
Pitch variation, speech pacing, and fluency — the vocal nuances that clinical screening values
Sentence complexity, grammatical patterns, and coherence across the conversation
All signals combined into a single, evidence-based prediction — with no black-box guesswork
Upload a file and get a full AI-powered analysis immediately — no waiting, no setup
Every result comes with SHAP explanations and counterfactual analysis — so you always know why
Train custom models on your own datasets with full control over features and algorithms
Recording guidelines
Conversation partner required. The recording must include the child speaking with a parent, guardian, or familiar adult — not the child alone.
Minimum 5 minutes. A longer, uninterrupted conversation gives the model enough turn-taking, pause, and topic data for a reliable result.
Keep it natural. Use everyday topics — activities, food, toys, stories. Avoid scripted or one-word-answer tasks.
Mix question types. Try open questions "What did you do today?", descriptive prompts "Tell me about your drawing", and follow-up questions "And then what happened?".
Drop audio file here
WAV, MP3, FLAC, OGG
At least 5 min · with a conversation partner
Acoustic Analysis
Loading visualization...