huben.com

What It Is

Songbird Speech is an AI clinical screening tool for speech and language concerns. Users record five standardized samples in the browser, including the Rainbow Passage, tongue twisters, sustained vowels, connected speech, and minimal pairs. They then receive a report across six major domains, which can be shared with a speech-language pathologist (SLP) via a secure link. The product clearly emphasizes that its output is a screening report, not a diagnosis.

Core Capabilities

Its AI stack consists of Claude AI, OpenAI Whisper, and browser-side DSP. Claude is used to assess six clinical domains: fluency, articulation, voice, aphasia indicators, language use, and motor speech. Whisper handles transcription and automatically identifies disfluency signals such as filler words, repetitions, and revisions. DSP extracts pause counts, speaking rate, pitch estimates, and energy variance directly in the browser. The final report includes a six-dimensional radar chart, urgency score, key findings, follow-up timeline, and referral recommendations.

Pricing and Ease of Use

The page states “Pay per assessment. No subscription,” but the patient pricing section also shows “Free per assessment,” so the actual pricing rules for patients are not entirely clear. Clinicians can join for free and receive shared assessments. The Members plan costs $9.99/month and provides access to the referral directory, the ability to analyze initial patient assessments, and the option to offer assessments to patients. No app download is required; the workflow of completing a questionnaire, recording in the browser, and viewing the report is claimed to take under 5 minutes, making the barrier to entry relatively low.

Pros, Cons, and Limitations

Its main strength is its focused use case: it directly addresses issues such as insufficient information before an initial consultation, difficulty finding an SLP, and barriers caused by geography or wait times. The shareable report also makes collaboration easier for clinicians, schools, or referral partners. The limitations are also clear: the main content does not provide clinical validation data, accuracy metrics, applicable age ranges, or condition boundaries. The claim of “clinical-grade reliability” is not backed by evidence. On privacy, it only mentions secure sharing links and that some features are processed in the browser, without explaining how audio, transcripts, and reports are stored, encrypted, or handled for healthcare compliance.

Who It’s For and Access from China

It is best suited for individuals with concerns about speech, articulation, fluency, voice, or language who want an initial screening, as well as SLPs who want a structured baseline before the first session. For users in China, the main content does not clarify support for Chinese speech, cross-border access, payment methods, or local compliance. Since it depends on Claude and Whisper, real-world network availability is uncertain, so china_access can only be rated as unknown. For Chinese-language use cases, it may still be better to prioritize assessment by local medical institutions or alternatives that support Chinese speech.