How VERA Works

AI Transparency & Technology Overview | Last Updated: March 13, 2026

VeraNeural, Inc. is committed to transparency about how VERA works, the technology it uses, and its limitations. This page explains what VERA is, how it generates responses, and what you should know as a user.

1. What VERA Is

VERA is:

• An AI-powered wellness companion that offers emotional support, self-reflection tools, and nervous system regulation guidance
• A conversational interface that uses neuroscience-informed frameworks to structure its responses
• A tool for personal exploration and emotional awareness, designed for daily use between professional appointments or as a standalone self-care resource
• Available on web and mobile platforms with text-based and voice-enabled interactions

2. What VERA Is Not

VERA is not:

• A medical device regulated by the FDA or any other medical device regulatory body
• A replacement for professional therapy, counseling, psychiatric care, or any medical treatment
• A diagnostic tool capable of assessing, diagnosing, or treating medical or mental health conditions
• A crisis intervention or emergency response system
• A clinical decision support system used by healthcare providers
• A tool monitored by licensed clinicians or mental health professionals in real time

3. How VERA Generates Responses

When you interact with VERA, your messages are processed through a multi-step pipeline:

1. When you send a message, it is transmitted securely to our servers
2. Your message is processed through VERA's behavioral frameworks, which add neuroscience-informed context and structure to the conversation
3. The structured prompt is sent to an external large language model (currently Anthropic Claude) which generates the text response
4. The response is returned to you through VERA's interface
5. Your conversation history is maintained within your account to provide continuity across sessions

Because VERA relies on probabilistic AI language models, responses are generated based on statistical patterns in language — not on clinical understanding, personal knowledge of your situation, or deterministic rules. This means responses may contain errors, inaccuracies, or interpretations that do not fully reflect your circumstances.

4. Behavioral Frameworks

What makes VERA different from a generic chatbot is the layer of neuroscience-informed behavioral frameworks that shape how it communicates. These frameworks are proprietary to VeraNeural and represent significant intellectual property.

Important: While these frameworks draw from established research, their application through an AI system has not been clinically validated. VERA's use of these frameworks does not constitute clinical practice and should not be interpreted as such.

5. External Technology

VERA uses the following external services and technologies to operate:

• Anthropic Claude: External large language model used to generate text responses. Anthropic processes your messages to generate responses but does not use them for model training under VeraNeural's data processing agreement.
• Clerk: Authentication provider that manages secure sign-in and account management.
• Supabase: Secure database infrastructure for storing account data and conversation history.
• Stripe: Payment processing for subscriptions. VeraNeural does not store credit card numbers directly.
• Text-to-Speech providers: Optional voice features may use services like ElevenLabs or OpenAI for speech synthesis.

6. Your Data and AI Training

• Your conversations are processed by external AI models (Anthropic Claude) solely to generate responses
• Your conversations are NOT used to train, fine-tune, or improve foundational AI language models
• Your conversation data is stored securely to maintain continuity within your account
• Anonymized, aggregated usage patterns may be analyzed to improve overall service quality
• VERA's behavioral frameworks are developed from published academic research and proprietary systems — not from individual user data
• You may request deletion of your conversation data at any time by contacting support@veraneural.com

7. Safety Positioning

VERA includes automated safety measures, but these have significant limitations. Our safety systems include:

• Automated keyword and pattern detection for potential crisis language
• Surfacing of crisis hotline numbers and emergency resources when triggered
• Clear disclaimers within the interface that VERA is not a crisis service

For a full list of crisis resources, visit our Safety & Crisis Resources page.

8. Questions and Feedback

We welcome questions and feedback about how VERA works. Transparency is fundamental to our mission, and we are committed to answering your questions honestly.