User agreement & AI disclaimers

What Crucible provides

Crucible is an educational and professional-development tool for aspiring and credentialed coaches working with the 2026 ICF Core Competencies. Every feature is powered by generative artificial intelligence and exists to develop your coaching craft.

• Transcript evaluation — you submit a coaching transcript (pasted, uploaded, or transcribed from audio) and the AI returns a structured developmental report mapping observations to ICF competencies with timestamped or turn-marked quotes, an indicative proficiency level, and suggested next-level moves.
• AI client practice — the AI role-plays a coaching client, either randomly generated or described by you, so you can practice coaching in a low-stakes environment. The client is fictional. Any resemblance to a real person is coincidental.
• MCC demonstration coaching — the AI models master-level coaching with you in the client chair, annotated with notes naming the competency in play. This is a demonstration of technique for learning purposes; it is not a professional coaching engagement.
• Spoken sessions — practice and demonstration sessions can be held hands-free by voice. Your microphone audio is captured in turns, transcribed, and the reply is spoken back with a synthetic voice. You can interrupt the AI voice at any time, and you can switch between talking and typing at any point.
• Transcripts and reports you can keep — every session transcript and evaluation report can be exported as a PDF to save or share from your phone or computer.
• Level calibration — you declare your credential level (Aspiring, ACC, PCC, MCC) and feedback is calibrated to that standard. If you are not credentialed, the AI offers an indicative assessment of the level your coaching currently demonstrates.

AI is a developmental tool only

All content in this platform is generated by large language models. It is offered solely as practice material and developmental feedback. It is not professional advice, not supervision, not mentor coaching hours, and not a substitute for a qualified human coach, mentor coach, or ICF assessor.

Output may be inaccurate

AI systems can be confidently wrong. They may misquote a transcript, invent timestamps, misapply a competency, produce inconsistent ratings between runs, or state something that is simply untrue. Assessments of proficiency level are indicative estimates produced by a model, never a determination of skill.

Always verify AI output against the transcript and against the official ICF materials before relying on it. Use your own professional judgement. You remain responsible for your coaching practice.

No credentialing or ICF affiliation

Crucible is independent and is not affiliated with, endorsed by, accredited by, or reviewed by the International Coaching Federation. Nothing produced here counts toward ICF credentialing, CCE units, mentor coaching requirements, coaching experience hours, or performance evaluation outcomes. A favourable evaluation here does not predict an ICF assessment result.

Coaching is not therapy

Coaching is a forward-focused, non-clinical partnership with individuals presumed to be creative, resourceful and whole. It does not diagnose, treat, or heal mental health conditions, trauma, substance use disorders, or psychological distress. This platform is not a mental health service, does not provide therapy or counselling, and must never be used as one — including in demonstration mode where the AI coaches you.

If you are experiencing psychological distress, seek a licensed mental health professional. In an emergency, contact local emergency services.

Safety flagging and its limits

The AI is instructed to interrupt a session and surface crisis resources when content suggests possible self-harm, suicidal ideation, harm to others, abuse, or acute crisis. This is an automated best-effort safeguard, not monitoring, not detection, and not intervention. It may miss serious situations and it may flag harmless ones. No human reviews these flags in real time and no emergency response is dispatched. Never rely on this platform in an emergency.

Confidentiality and what you submit

Transcripts and session content are stored in your account so you can return to them, and are transmitted to third-party AI model providers for processing. Do not submit content you are not authorised to share. Obtain your client's informed consent before uploading any recording or transcript of a real session, and anonymise identifying details — names, employers, locations, health information — wherever possible.

Do not submit special-category personal data, protected health information, or anything covered by a confidentiality agreement that does not permit AI processing.

Voice sessions, microphone and synthetic speech

Voice mode only works while you have granted your browser microphone permission and have voice mode switched on. While it is on, the platform listens continuously, captures what you say in turns, and sends that audio to a third-party AI provider for transcription. Only the resulting text is stored with your session; the captured audio is not retained as a recording by the platform.

Be mindful of your surroundings. Anyone speaking near your device may be captured and transcribed. Do not use voice mode where confidential or third-party conversations can be overheard, and switch to typing whenever you are unsure.

All spoken replies are synthetic speech generated by AI. Voices are chosen to suit a fictional persona and do not represent, imitate or clone any real person. Voice detection is imperfect: it may cut you off, miss an interruption, or mis-transcribe what you said, and any resulting error will flow through to the feedback you receive.

Exports you download

Transcripts and evaluation reports can be exported as PDF files. Once a file leaves the platform it is entirely under your control and outside the protections described here. Store, share and dispose of exported client material in line with your own confidentiality obligations.

Acceptable use

You agree not to use the platform to:

• Present AI-generated feedback as an official ICF assessment or as human evaluation.
• Provide therapy, counselling, medical, legal or financial advice to anyone using output from this platform.
• Upload real client material without informed consent.
• Attempt to elicit harmful, illegal, or abusive content, or circumvent the platform's safety instructions.
• Resell, scrape, or redistribute platform output as your own assessment product.

Sandbox status, availability and accounts

The platform currently operates as a sandbox environment for evaluation purposes. Accounts, stored sessions, and evaluations may be reset, changed, or removed without notice, and paid access may be introduced later. Service is provided on an as-is, as-available basis without warranties of any kind.

Limitation of liability

To the maximum extent permitted by law, the operators of Crucible are not liable for any decision, action, coaching outcome, client outcome, credentialing result, or loss arising from your use of, or reliance on, AI-generated content in this platform. You use it at your own professional risk.