← Back to homeHow we use AI responsibly
ParcAI processes sensitive personal information — health, finances, family, beliefs. This page explains exactly what we do with your data, which models we use, how our autonomous Agent operates, and the limits of what our AI can and cannot do.
Last updated: June 16, 2026
1. Which AI models process your data
ParcAI uses a proprietary AI gateway (LiteLLM) that routes your messages to the most appropriate models by task type:
| Provider | Use | Data policy |
|---|---|---|
| Gemini Flash 2.0 (Google) | General responses and fast tasks | Google AI Policy → |
| Claude Sonnet 4 (Anthropic) | Sensitive domains (health, family, faith, legal) + Agent orchestration | Anthropic Policy → |
Neither provider uses your conversations to train AI models when accessed via API (unlike free web usage). We review the privacy policies of each provider before routing any data.
2. What we store about you
Contextual memory
Facts you share and authorize ParcAI to remember (e.g. “my son is 8 years old”, “I work as a nurse”). Stored in your profile and used to personalize future responses. View and delete everything in Account → My Data.
Conversation history
Conversation threads stored so you can pick up where you left off. Not used to train models. You can delete any thread individually.
Active missions
Action plans created with the agents. Visible in Mission Control and can be paused or deleted at any time.
Agent action logs
Records of tasks the Agent executed autonomously (mission creation, memory updates, multi-step flows). Retained for 30 days so you can review and correct Agent activities.
Usage data (anonymized)
Aggregated metrics like "which agent was most used" or "how many missions were created", without personal identification. Used to improve the product.
3. How the Agent works — autonomy and limits
The ParcAI Agent can execute multi-step tasks autonomously within the platform. Here is exactly what it can and cannot do without your explicit confirmation:
✓ Can do autonomously
- • Create and update Missions and action plans
- • Organize and summarize contextual memory
- • Synthesize information across conversation threads
- • Set internal reminders and structured tasks
- • Suggest next steps and prioritize your agenda
- • Coordinate multi-domain analysis (finance + health + career)
✗ Requires your explicit confirmation
- • Sending messages to third parties on your behalf
- • Accessing or modifying external integrations (Google Drive, Calendar)
- • Financial transactions or payments
- • Publishing content to social media or external platforms
- • Deleting data irreversibly
- • Any action with permanent external consequences
You can review, pause, or roll back any Agent action in Mission Control at any time. The Agent logs all autonomous actions for 30 days so you always have full visibility and control.
4. Agent limits — what AI does not replace
Our agents are reflection and organization partners, not substitutes for professionals. See the explicit limits for each sensitive domain:
Health & Wellness
Does not replace medical diagnosis, medication prescription, or guidance from a healthcare professional. In emergencies, call local emergency services.
Spiritual Partner
Personal reflection support only. Does not represent official religious guidance from any institution.
Family Partner
Support for family reflections. Situations of domestic violence or abuse: contact emergency services or hotlines.
Elderly Care Partner
Organizational support for caregivers. Does not replace geriatric evaluation or in-person medical care.
Legal Partner
General information about rights only. For specific legal matters, consult a licensed attorney.
Social Partner
If you are in intense emotional distress or having thoughts of self-harm, please contact a crisis line immediately.
Finance Partner
Does not constitute professional financial or investment advice. Consult a licensed financial advisor for investment decisions.
5. What ParcAI will never do
- Sell or share your conversations with advertisers
- Use your personal data to train AI models
- Take autonomous decisions with external consequences without your approval
- Display ads based on the content of your conversations
- Share Squad members' data without individual consent
- Allow the Agent to make financial transactions without explicit confirmation
- Store your data beyond the retention periods defined in our Privacy Policy
6. Your rights over your data
In compliance with LGPD (Law No. 13,709/2018), you can at any time:
- View your memoryAccount → My Data
- Export all your dataAccount → My Data → Export
- Delete your memoryAccount → My Data → Delete
- Review Agent action logsMission Control
- Delete your accountEmail contato@parcai.com.br
For any request: contato@parcai.com.br — we respond within 15 business days.
Found something wrong?
If an agent or the autonomous Agent gave a response you consider harmful, inaccurate, or inappropriate, let us know. We take this seriously.
Report via emailOr use the /feedback command directly in the chat.