How to use Auven

Create your account

Sign up with your institutional or personal email. Select your role — student or supervisor — and complete a short onboarding flow to personalise your experience.

Set up your first session

Choose a client profile and launch your first voice session in under two minutes.

Understanding client profiles

Learn how Auven models realistic clients with distinct psychological profiles, attachment styles, and presenting concerns grounded in clinical literature.

Voice mode setup

Grant microphone permissions, check your audio input device, and run a quick sound test before going live with a voice session.

Generating AI clients

Use the client generator to define demographics, presenting issues, and personality traits. The AI builds a consistent, clinically grounded persona from your inputs.

Custom scenarios

Write a brief scenario description and let Auven turn it into a fully playable client complete with backstory, session goals, and resistance patterns.

Difficulty settings

Adjust how much the client resists rapport-building, how forthcoming they are with information, and how quickly they escalate emotionally during sessions.

Cognitive diversity options

Configure profiles that include ADHD, autism spectrum traits, trauma responses, or psychosis to prepare for a wider range of clinical encounters.

Starting a practice session

Select a saved client or generate a new one and press Start Session. A brief intake summary is shown before the first exchange begins.

Voice mode

Voice mode uses your microphone for real-time spoken conversation with the AI client. Grant microphone permissions when prompted before starting.

Session controls

Pause, resume, or end a session at any time. Use the hint button for a suggested response if you feel stuck, without breaking session flow.

Ending and reviewing

End the session to unlock your score report, full transcript, and turn-by-turn feedback on empathy, technique use, and client alliance.

Session scores explained

Each session produces an overall score out of 100, combining empathy accuracy, therapeutic technique, pacing, and client-reported alliance.

Alliance score

The working alliance metric reflects how well the client feels heard and understood. It updates turn by turn and is visible in full after the session ends.

EQ metrics

Emotional intelligence sub-scores cover active listening, non-judgmental language, validation quality, and reflective accuracy — each linked to specific transcript moments.

Progress over time

The progress dashboard shows score trends across sessions with breakdowns per client type and skill category so you can identify patterns in your development.

Creating student accounts

Invite students by email from your supervisor dashboard. Each student account is linked to your cohort and you gain read access to their session analytics automatically.

Reviewing sessions

Open any student session to view a full annotated transcript, score breakdown, and AI-flagged moments suitable for discussion in live supervision.

Custom scenario libraries

Build a library of approved scenarios that all students in your cohort can access. Scenarios can be locked to specific difficulty tiers or topic areas.

Progress reports

Export PDF or CSV progress reports per student or cohort, covering session frequency, average scores, and skill-level trajectories over time.

Authentication

All API requests require a Bearer token obtained via /api/auth/token. Tokens expire after 24 hours and can be refreshed without re-authentication.

Endpoints overview

The REST API exposes endpoints for sessions, client profiles, analytics, and user management. A full OpenAPI spec is available on the developers page.

Webhooks

Subscribe to session.completed, score.updated, and user.created events. Payloads are signed with HMAC-SHA256 so you can verify authenticity server-side.

SDK guides

Official TypeScript and Python SDKs wrap the REST API and include helpers for streaming voice sessions over WebSocket connections.