Eden includes powerful AI features — transcription, smart search, AI chat, and auto-tagging. This article explains exactly how your data is used (and not used) when you interact with these features.
The Short Version
We do NOT use your content to train AI models. Your files, videos, notes, and conversations are never fed into training datasets. The AI features work for you alone.
However, when you use AI chat, your content is processed by third-party AI providers to generate responses. Here's how that works and what it means for your privacy.
What AI Features Does Eden Offer?
Eden uses AI in several ways:
Feature | What It Does | When It Runs |
|---|---|---|
Automatic Transcription | Converts speech in videos and audio to searchable text | When you upload a video or audio file |
Smart Search | Finds content based on meaning, not just keywords | When you search your workspace |
Frame Analysis | Analyzes what's in each video frame | When videos are processed |
Auto-tagging | Generates tags and descriptions for your content | During file processing |
AI Chat | Answers questions about your content using AI models | When you send a message in an AI chat |
Your Content Is NOT Used for Training
This is important, so we'll be direct:
Eden does not use your content to train or improve AI models
Your files, notes, transcripts, and conversations are not added to any training datasets
No one else benefits from your content — it's processed for you alone
This applies to all AI features, including transcription, search, and chat.
How AI Chat Works
When you chat with AI on a Canvas, here's what happens:
You send a message — Your question plus any connected content (transcripts, documents, etc.) is prepared for the AI
Data is sent to the AI provider — Your content travels to one of our AI partners over an encrypted connection
The AI generates a response — The provider processes your content and returns an answer
You see the result — The response appears in your chat
The AI provider processes your content to generate a response, then the data is not retained for training purposes.
Our AI Providers
Eden works with multiple AI providers so you can choose the best model for your task:
Provider | Models |
|---|---|
Anthropic | Claude (Haiku, Sonnet, Opus) |
OpenAI | GPT-4o, GPT-5, GPT-5.1 |
Gemini 2.5 Flash, Gemini 2.5 Pro, Gemini 3 | |
xAI | Grok 4, Grok 4 Fast |
All of these providers have committed to not using data sent through their APIs for model training. Your content is processed to generate a response, then discarded.
What Each Provider Promises
We've selected AI partners with strong privacy commitments:
Anthropic (Claude): Data sent via API is not used to train models. Content is processed and then deleted.
OpenAI: API data is not used for training by default. OpenAI's data usage policy for API customers differs from their consumer products.
Google (Gemini): API data is processed for your request only and not used for model improvement.
xAI: Similar API-specific privacy terms apply.
You can review each provider's policies directly on their websites for the most current information.
Data Flow Summary
Here's a visual overview of how your data moves through Eden's AI features:
Transcription & Processing
AI Chat
What We Store vs. What Providers See
Data | Stored by Eden | Seen by AI Providers | Used for Training |
|---|---|---|---|
Your original files | ✅ Yes | ❌ No (only transcripts/text) | ❌ No |
Transcripts | ✅ Yes | ✅ Yes (during processing) | ❌ No |
AI chat messages | ✅ Yes | ✅ Yes (during processing) | ❌ No |
AI responses | ✅ Yes | Generated by them | ❌ No |
Search queries | ✅ Yes | ✅ Yes (for semantic search) | ❌ No |
Controlling AI Features
You have control over how AI interacts with your content:
Choose your model: Different models have different capabilities. You can switch models in any AI chat.
Connect selectively: On Canvas, AI only sees content you explicitly connect to a chat node. Unconnected items aren't sent to the AI.
Use folders strategically: Keep sensitive content in separate folders and only connect what you need.
Sensitive Content Considerations
While our AI providers don't use your content for training, you should still be thoughtful about what you process through AI features:
Highly confidential business information: Consider whether AI chat is necessary
Personal information about others: Be aware that this data passes through third-party systems
Regulated data (HIPAA, etc.): Check with your compliance team before processing
For most creative and professional work, Eden's AI features are safe to use. But if you're handling extremely sensitive data, consider which features you use.
Enterprise Considerations
If you're evaluating Eden for enterprise use:
All AI providers have enterprise-grade API agreements
Data processing agreements are in place with each provider
SOC 2 Type II certification is in progress
We can provide additional security documentation upon request
Contact us at support@eden.so for enterprise security discussions.
Common Questions
Can I use Eden without AI features?
Eden's AI features are integral to how the product works — transcription and smart search happen automatically. However, AI chat is optional; you don't have to use it.
Does the AI "remember" my previous conversations?
AI models don't retain memory between sessions. Each conversation starts fresh. Within a single chat session, the AI can reference earlier messages in that conversation.
What if an AI provider changes their policies?
We monitor our providers' policies and would notify users of any significant changes. We select partners with strong, consistent privacy commitments.
Can I request deletion of data sent to AI providers?
Data sent to AI providers for processing is transient — it's used to generate a response and not retained. There's nothing to delete because nothing is stored on their end.
Does using AI features cost extra?
AI features use your monthly request allowance. See Understanding AI Requests for details on how requests work.
