> ## Documentation Index
> Fetch the complete documentation index at: https://docs.enconvo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Privacy & Security

> How Enconvo handles your data and keeps your information secure

## Overview

EnConvo is designed with privacy as a core principle. Your data stays on your Mac by default, API keys are stored securely in the macOS Keychain, and you have full control over when data leaves your device.

## Data Storage

### Local-First Architecture

All EnConvo data is stored locally on your Mac:

| Data Type        | Storage Location               | Format                                |
| ---------------- | ------------------------------ | ------------------------------------- |
| Conversations    | `~/.config/enconvo/`           | SQLite database                       |
| Knowledge bases  | `~/.config/enconvo/`           | LanceDB (vectors) + SQLite (metadata) |
| Extension data   | `~/.config/enconvo/extension/` | Files and SQLite                      |
| Preferences      | `~/.config/enconvo/`           | JSON files                            |
| Cache            | `~/.config/enconvo/cache/`     | Temporary files                       |
| Audio recordings | `~/.config/enconvo/`           | Audio files (MP3, WAV)                |

<Tip>
  All your conversations, knowledge bases, and settings are stored under `~/.config/enconvo/`. Back up this directory to preserve your data.
</Tip>

### No Cloud Sync by Default

EnConvo does not automatically sync data to any cloud service. Your conversations, knowledge bases, and preferences remain exclusively on your local machine unless you explicitly use a cloud-based feature.

## API Key Security

### macOS Keychain Storage

All API keys and credentials are stored in the macOS Keychain, the system's built-in encrypted credential store:

* Keys are encrypted at rest using your macOS login credentials
* Keys are never written to plain-text files
* Keys are accessible only to EnConvo and only while your Mac is unlocked
* Keys are never logged or included in crash reports

### Key Management

<Steps>
  <Step title="Adding a key">
    Settings → Key Management → Select Provider → Enter Key
    The key is immediately encrypted and stored in Keychain.
  </Step>

  <Step title="Validating a key">
    Click "Validate" to test the key with the provider's API.
    Only a minimal test request is sent.
  </Step>

  <Step title="Removing a key">
    Delete the key from Settings. It is immediately removed from Keychain.
  </Step>
</Steps>

## Network Requests

### When Data Leaves Your Device

EnConvo only sends data over the network in these specific situations:

| Scenario                   | What Is Sent                           | Destination                                       |
| -------------------------- | -------------------------------------- | ------------------------------------------------- |
| **Cloud AI chat**          | Your messages and context              | AI provider API (OpenAI, Anthropic, Google, etc.) |
| **Web search**             | Search queries                         | Search provider (Tavily, Bing, Google, etc.)      |
| **TTS/STT**                | Text or audio                          | TTS/STT provider API                              |
| **Extension store**        | Extension metadata requests            | enconvo.com                                       |
| **License validation**     | License key                            | enconvo.com                                       |
| **Cloud image generation** | Text prompts                           | Image generation provider                         |
| **Email integration**      | Email content (when using Gmail tools) | Gmail API                                         |

### What Is Never Sent

* Your local files (unless you explicitly use them as context)
* Your screen content (unless you explicitly use screenshot context)
* Your clipboard content (unless you explicitly include it)
* Your browsing history
* Application usage data
* Other application content

<Note>
  When using cloud AI providers, your messages are sent to their APIs and subject to their respective privacy policies. Review each provider's data handling practices.
</Note>

## Local AI: Fully Offline Mode

For maximum privacy, use local AI models that run entirely on your Mac:

### Local LLM Options

| Provider      | Setup                              | Privacy Level                      |
| ------------- | ---------------------------------- | ---------------------------------- |
| **Ollama**    | Install Ollama, download models    | Full offline -- zero network calls |
| **MLX**       | Built-in Apple Silicon models      | Full offline -- zero network calls |
| **LM Studio** | Install LM Studio, download models | Full offline -- zero network calls |

### Local TTS/STT Options

| Feature | Local Provider                     | Notes                                       |
| ------- | ---------------------------------- | ------------------------------------------- |
| **TTS** | macOS System TTS                   | Built into macOS, no download needed        |
| **STT** | NVIDIA Parakeet, Qwen ASR, Whisper | Requires model download, then fully offline |
| **OCR** | Apple Native Vision OCR            | Built into macOS, fully offline             |

### Setting Up Full Offline Mode

<Steps>
  <Step title="Install Ollama">
    Download from ollama.com and install a model:

    ```bash theme={null}
    ollama pull llama3.2
    ```
  </Step>

  <Step title="Configure EnConvo">
    Settings → AI Models → Default Model → Select Ollama
  </Step>

  <Step title="Use local providers">
    Set local providers for TTS (macOS System), STT (Parakeet), and OCR (Apple Native)
  </Step>

  <Step title="Disable cloud features">
    Avoid selecting cloud-based providers in any extension settings
  </Step>
</Steps>

<Tip>
  With local models configured, you can disconnect from the internet entirely and EnConvo will continue to work. This is ideal for working with sensitive documents.
</Tip>

## Extension Security

### Extension Sandboxing

Extensions run in isolated Node.js Worker threads with these constraints:

* Each command executes in its own Worker thread
* Workers have a 30-minute idle timeout and are automatically cleaned up
* Extensions can only access the file system through defined APIs
* Extensions communicate with the host app through a structured JSON-RPC protocol

### Extension Source Verification

* All official extensions are open source on [GitHub](https://github.com/enconvo)
* Extensions from the store are reviewed before publication
* You can inspect any extension's source code before installing
* Side-loaded extensions (from folder or ZIP) run with the same permissions as store extensions

### Extension Permissions

Extensions declare their capabilities in `package.json`. When an extension tries to use a system feature (file access, network, clipboard), it goes through EnConvo's API layer which mediates access.

## Telemetry & Analytics

### Anonymous Usage Statistics

EnConvo collects minimal, anonymous usage statistics to improve the product:

* Feature usage frequency (which features are used, not what content)
* Error and crash reports (stack traces only, no user content)
* App version and macOS version

### What Is Not Collected

* Conversation content
* AI prompts or responses
* File contents
* Screen captures
* API keys or credentials
* Personal information

### Opting Out

You can disable telemetry entirely:

Settings → General → Analytics → Toggle Off

When disabled, EnConvo sends zero telemetry data.

## Security Best Practices

<AccordionGroup>
  <Accordion title="Use local models for sensitive work">
    When working with confidential documents, legal materials, or proprietary code, configure local LLM providers (Ollama, MLX) to ensure nothing leaves your device.
  </Accordion>

  <Accordion title="Review extension permissions">
    Before installing third-party extensions, review their source code on GitHub. Check what APIs they use and what network requests they make.
  </Accordion>

  <Accordion title="Rotate API keys regularly">
    Periodically rotate your AI provider API keys. Remove old keys from EnConvo's Key Management and add fresh ones.
  </Accordion>

  <Accordion title="Keep EnConvo updated">
    Updates include security patches. Enable auto-updates or check for updates regularly.
  </Accordion>

  <Accordion title="Back up your data">
    Regularly back up `~/.config/enconvo/` to preserve your knowledge bases, conversations, and settings. Use Time Machine or your preferred backup tool.
  </Accordion>

  <Accordion title="Lock your Mac">
    Since API keys are protected by your macOS login, always lock your Mac when stepping away. Keychain access requires an unlocked session.
  </Accordion>
</AccordionGroup>

## Data Deletion

### Deleting Conversations

* Delete individual conversations from the chat list
* Clear all conversations from Settings → Advanced → Clear Data

### Deleting Knowledge Bases

* Delete individual knowledge bases from the Knowledge Base manager
* All associated vector data and metadata are removed

### Complete Data Removal

To remove all EnConvo data from your Mac:

```bash theme={null}
# Remove all EnConvo data
rm -rf ~/.config/enconvo/
rm -rf ~/.config/Enconvo/

# Remove API keys from Keychain
# Open Keychain Access app → Search "enconvo" → Delete entries
```

<Warning>
  This permanently deletes all conversations, knowledge bases, extensions, and preferences. This action cannot be undone.
</Warning>

## Next Steps

<CardGroup cols={2}>
  <Card title="Settings Reference" icon="gear" href="/configuration/settings-reference">
    Configure all EnConvo settings
  </Card>

  <Card title="Local LLM" icon="computer" href="/ai/local-llm">
    Set up fully local AI models
  </Card>

  <Card title="Knowledge Base" icon="database" href="/ai/knowledge-base">
    Learn about knowledge base data storage
  </Card>

  <Card title="Troubleshooting" icon="wrench" href="/getting-started/troubleshooting">
    Fix common issues
  </Card>
</CardGroup>