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

# References & Sources

> Track and view data sources used in conversations

## Overview

References and sources in Pulze provide transparency into the data and information used by AI assistants in their responses. See exactly which documents, files, and data sources were consulted to generate each answer.

## Viewing Sources

### Source Display

When an assistant uses data sources, you'll see:

* **Source Icon** - Visual indicator that sources were used
* **Source Count** - Number of sources referenced
* **Source Preview** - Quick glimpse of source content
* **Full Source Details** - Complete source information

### Per-Message Sources

Sources are displayed:

* Attached to specific assistant responses
* Expandable for detailed view
* Clickable to view source content
* Grouped by type (documents, custom data, web sources)

## Source Types

### Documents

References to documents in your spaces:

* Document title
* Relevant excerpt
* Location within document
* Last modified date

### Custom Data

Files and data uploaded to Pulze:

* File name
* Relevant section
* Upload date
* File type

### Web Sources

External web pages:

* Page title
* URL
* Excerpt used
* Access date

### Datasets

Evaluation and test datasets:

* Dataset name
* Specific data points
* Entry details

## Source Details

### Viewing Full Content

Click on any source to see:

* Complete source text
* Context around the referenced section
* Source metadata
* Related sources

### Source Preview

Hover over sources for quick preview:

* Excerpt of relevant content
* Source type
* Relevance score (if available)

## Source Attribution

### Why Sources Matter

Source tracking provides:

* **Transparency** - See where information comes from
* **Verification** - Validate assistant responses
* **Traceability** - Track data lineage
* **Compliance** - Meet regulatory requirements
* **Trust** - Build confidence in AI responses

### Citation Format

Sources are cited with:

* Source identifier
* Reference number
* Link to original content
* Context information

## Source Relevance

### Relevance Scoring

Sources may display relevance indicators:

* How closely source matches the query
* Confidence in source selection
* Quality of source content
* Recency of information

### Primary vs. Secondary Sources

Distinguish between:

* **Primary Sources** - Directly referenced in response
* **Secondary Sources** - Consulted but not directly quoted
* **Background Sources** - General context

## Managing Sources

### Excluding Sources

Control which sources are used:

* Exclude specific sources from search
* Prioritize certain data types
* Configure source preferences
* Manage data associations

### Source Access

Ensure sources are accessible:

* Verify file permissions
* Check data associations
* Confirm source availability
* Review access controls

## Source Search

### Finding Specific Sources

Search for sources by:

* File name
* Content keywords
* Date range
* Source type
* Space association

### Source Filtering

Filter displayed sources:

* By type (document, custom data, web)
* By relevance
* By date
* By space

## Integration with Data

### Linked to Data Section

Sources connect to:

* Data management interface
* Document repository
* Custom data uploads
* Dataset collections

### Data Updates

When source data changes:

* Assistant has access to latest version
* Previous conversation sources preserved
* Transparency in version used

## Privacy and Security

### Source Access Control

Source visibility respects:

* Space permissions
* Member access levels
* Data classification
* Privacy settings

### Sensitive Information

Be aware of:

* What data sources contain
* Who can see referenced sources
* Compliance requirements
* Data governance policies

## Source Quality

### Evaluating Sources

Consider source quality:

* Accuracy of information
* Recency of data
* Relevance to query
* Authority of source

### Improving Source Quality

Enhance source quality by:

* Keeping data updated
* Removing outdated sources
* Adding authoritative sources
* Organizing data effectively

## Best Practices

<Tip>
  **Verify Important Information**: Always check sources for critical decisions or sensitive information.
</Tip>

<Tip>
  **Keep Data Current**: Regularly update your data sources to ensure assistant has latest information.
</Tip>

<Tip>
  **Review Source Usage**: Periodically check which sources are being referenced to ensure appropriateness.
</Tip>

<Tip>
  **Organize Sources**: Use clear naming and organization for data sources to improve discoverability.
</Tip>

## Source Analytics

### Usage Tracking

Monitor which sources are:

* Most frequently referenced
* Providing best results
* Underutilized
* Causing issues

### Insights

Gain insights from source usage:

* Popular information needs
* Knowledge gaps
* Content effectiveness
* Assistant behavior patterns

## Troubleshooting

### Sources Not Appearing

If sources aren't shown:

1. Verify assistant has access to data
2. Check space data associations
3. Confirm files are properly indexed
4. Review assistant configuration

### Irrelevant Sources

If wrong sources are used:

1. Review data organization
2. Improve file naming and metadata
3. Use labels to categorize data
4. Adjust assistant instructions

### Source Access Errors

If sources can't be accessed:

1. Check file permissions
2. Verify data still exists
3. Confirm space associations
4. Review member access rights

## Citation Export

### Exporting References

Export source citations:

* Include in conversation export
* Generate reference list
* Create bibliography
* Document sources used

### Formats

Export in various formats:

* Plain text
* Markdown
* Citation styles (APA, MLA, etc.)
* Custom formats

## Compliance and Audit

### Audit Trail

Source tracking provides:

* Record of information used
* Timestamp of access
* Member who initiated query
* Complete traceability

### Regulatory Compliance

Meet requirements for:

* Data lineage
* Information governance
* Audit requirements
* Transparency mandates

## Advanced Features

### Source Ranking

Configure how sources are prioritized:

* By relevance
* By recency
* By authority
* Custom ranking criteria

### Source Filtering

Set up filters to:

* Include/exclude sources
* Prioritize certain types
* Limit by date range
* Constrain by space

### Smart Source Selection

Assistant intelligently selects sources based on:

* Query context
* Historical usage
* Source quality
* Relevance scores

<CardGroup cols={2}>
  <Card title="Chat Interface" icon="messages" href="/pulze/conversations/chat-interface">
    Back to Chat Interface
  </Card>

  <Card title="Message History" icon="clock-rotate-left" href="/pulze/conversations/message-history">
    View Message History
  </Card>

  <Card title="Data Management" icon="database" href="/pulze/data/overview">
    Manage Data Sources
  </Card>
</CardGroup>