Usage

With LegalEase running, you can start managing cases, processing documents, running transcriptions, and searching across your content. This guide covers the most common workflows.

Create a Case

Cases are the primary organizational unit in LegalEase. All documents and transcripts belong to a case.

1. Open the Cases page from the sidebar
2. Click New Case
3. Fill in the required fields:
   • Name - Descriptive title for the case
   • Case Number - Your internal reference number
   • Client - Client name or identifier
   • Matter Type - Optional category (litigation, contract, etc.)
4. Click Create to save

Cases start in active status. You can archive cases when work is complete.

Upload Documents

LegalEase processes PDFs and other documents through Docling for text extraction, chunking, and vector indexing.

1. Navigate to a case or the Documents page
2. Click Upload and select files (PDF, DOCX, images)
3. Documents enter processing status while:
   • Docling extracts text, tables, and structure
   • Pages are chunked into searchable segments
   • Vector embeddings are generated and indexed in Qdrant
   • Page renders are created for the PDF viewer
4. Once complete, status changes to completed

Document Viewer

Click any document to open the viewer:

• Navigate pages with thumbnails or keyboard arrows
• Search hits are highlighted with bounding boxes from the original extraction
• Sidebar shows extracted entities and metadata
• Toggle between text and visual layers

Transcribe Audio/Video

Upload audio or video files to generate AI-powered transcripts with speaker diarization.

1. Navigate to Transcripts or a case detail page
2. Click Upload Recording
3. Select your audio/video file (MP3, WAV, MP4, etc.)
4. The file uploads to Firebase Storage and processing begins:
   • Gemini 2.5 Flash (or Chirp 3) transcribes the audio
   • Speakers are identified and labeled
   • Speaker names are inferred from conversational context
   • Timestamps are assigned to each segment
5. Once complete, view the full transcript

Transcript Viewer

The transcript page provides rich interaction:

• Audio player - Play/pause with waveform visualization (in progress)
• Segment navigation - Click any segment to jump to that timestamp
• Speaker filtering - Show only specific speakers
• Edit speakers - Rename speakers for clarity
• Key moments - AI-identified important points in the conversation
• Summary - Executive summary, topics, and action items

Summarization

Transcripts are automatically summarized using Gemini:

• Executive summary - 1-2 paragraph overview
• Key moments - Timestamped highlights with importance ratings
• Action items - Follow-up tasks mentioned in the conversation
• Topics - Main subjects discussed
• Entities - People, organizations, locations, and dates mentioned

Search Across Content

LegalEase provides hybrid search combining semantic understanding with keyword matching.

1. Open the Search page
2. Enter your query - natural language works best
3. Optionally filter by:
   • Case - Limit to specific cases
   • Document type - Documents, transcripts, or both
   • Search mode - Hybrid (default), semantic-only, or keyword-only
4. Results show:
   • Matching text with context
   • Source document/transcript
   • Relevance score
   • Case assignment

Click any result to navigate directly to that location in the document or transcript.

Search Tips

• Use natural language: "What did the witness say about the contract date?"
• Combine keywords for precision: "Smith deposition damages"
• Filter by case when working on a specific matter
• Toggle search modes to compare results

Case Management

Archive Cases

When a case is complete:

1. Open the case detail page
2. Click the Archive button
3. Archived cases are hidden from the default list but remain searchable

Restore Archived Cases

1. Use the filter to show archived cases
2. Open the case and click Restore

Case Statistics

Each case shows:

• Total documents and their processing status
• Total transcripts and duration
• Recent activity

Export Content

Export transcripts in various formats (feature in progress):

• DOCX - Word document with speaker labels and timestamps
• SRT - Subtitle format for video players
• VTT - Web video text tracks
• TXT - Plain text transcript
• JSON - Structured data with all metadata

Monitor Background Jobs

Processing happens asynchronously via Firebase Cloud Functions:

• Document status - Watch the document card for status updates
• Transcript status - Real-time updates as transcription progresses
• Function logs - Check Firebase Emulator UI (http://localhost:4000) for detailed logs

Firestore triggers ensure status updates appear in real-time without page refresh.