​

2025 Changelog

​

June 2025

​

v1.8.0 - June 16, 2025

• NEW: Official Python SDK (verbalisai-sdk)
  • Full async/await support with asyncio
  • Automatic retry logic with exponential backoff
  • Comprehensive error handling and logging
  • Type hints for better IDE support
• NEW: Official JavaScript SDK (@verbalisai/sdk)
  • Full TypeScript support with complete type definitions
  • Works in Node.js, React, Next.js, and browser environments
  • Modern ESM and CommonJS module support
  • Built-in request/response interceptors

• NEW: Complete SDK documentation with examples
• NEW: Interactive code samples for all endpoints
• NEW: Framework-specific integration guides
• IMPROVED: API reference with better examples and error codes

• IMPROVED: 40% faster processing for mini model
• IMPROVED: 25% reduction in API response times
• IMPROVED: Enhanced file upload speeds for files >100MB
• FIXED: Memory optimization for long audio files

​

v1.7.2 - June 10, 2025

• NEW: Enhanced PII redaction policies
  • Added healthcare-specific PII detection (medical IDs, patient names)
  • Added financial PII detection (account numbers, routing numbers)
  • Custom PII substitution patterns
• NEW: Advanced entity detection
  • Support for custom entity types
  • Improved accuracy for person names and locations
  • Added confidence scores for detected entities

• IMPROVED: 99.9% webhook delivery reliability
• NEW: Automatic retry with exponential backoff (up to 5 attempts)
• NEW: Webhook signature verification for security
• NEW: Webhook testing tools in dashboard

​

v1.7.1 - June 3, 2025

• FIXED: Rate limiting headers now correctly reflect remaining quota
• FIXED: File upload timeout issues for files >500MB
• IMPROVED: Better error messages for invalid audio formats
• IMPROVED: Reduced memory usage during transcription processing

​

May 2025

​

v1.7.0 - May 28, 2025

• NEW: nano model for ultra-fast English transcriptions
  • 3x faster than mini model
  • Optimized for real-time applications
  • English-only support with 95%+ accuracy
• IMPROVED: pro model accuracy improvements
  • 15% better accuracy across all languages
  • Enhanced handling of accented speech
  • Better performance with background noise
• IMPROVED: mini model optimizations
  • 20% faster processing time
  • Better handling of multiple speakers
  • Improved timestamp accuracy

​

v1.6.5 - May 21, 2025

• NEW: Advanced usage analytics dashboard
  • Daily, weekly, monthly usage breakdowns
  • Cost analysis and projections
  • Audio format and model usage statistics
  • Export capabilities for usage data
• NEW: Real-time transcription monitoring
  • Live status updates for processing transcriptions
  • Queue position and estimated completion time
  • Historical processing time trends
• IMPROVED: API key management
  • Scoped permissions (read-only, write-only, admin)
  • API key usage tracking and analytics
  • Automatic key rotation capabilities

​

v1.6.0 - May 14, 2025

• NEW: OAuth 2.0 support for enterprise customers
  • PKCE flow for single-page applications
  • Refresh token rotation
  • Scope-based access control
• NEW: IP whitelisting for API keys
  • Restrict API key usage to specific IP ranges
  • Support for CIDR notation
  • Audit logs for IP-based access attempts
• IMPROVED: Enhanced audit logging
  • Detailed logs for all API operations
  • Retention options (30, 90, 365 days)
  • Export capabilities for compliance

​

v1.5.8 - May 7, 2025

• NEW: Advanced file organization
  • Folder structure for uploaded files
  • Tagging system for better organization
  • Bulk operations (move, delete, tag)
• NEW: Automatic file cleanup policies
  • Configure automatic deletion after N days
  • Archive old files to cold storage
  • Lifecycle management rules
• IMPROVED: File upload reliability
  • Resumable uploads for large files
  • Better error handling for network interruptions
  • Progress tracking for uploads

​

April 2025

​

v1.5.5 - April 30, 2025

• IMPROVED: Better handling of technical terminology
  • Enhanced medical and legal vocabulary
  • Improved recognition of product names and brands
  • Better handling of acronyms and abbreviations
• IMPROVED: Multi-language support
  • Added support for 5 new languages (Hindi, Arabic, Hebrew, Thai, Vietnamese)
  • Better code-switching detection
  • Improved accuracy for non-native speakers

​

v1.5.0 - April 16, 2025

• BREAKING: Updated authentication header format
  • Old: Authorization: YOUR_API_KEY
  • New: Authorization: Bearer YOUR_API_KEY
  • Migration guide: /changelog/migration/v2-to-v3
• BREAKING: Webhook payload format updated
  • Standardized payload structure across all events
  • Added timestamp and version fields
  • Migration guide available in dashboard

• NEW: Batch transcription API
  • Process up to 100 files in a single request
  • Automatic queue management
  • Bulk status checking and results retrieval
• NEW: Custom vocabulary support
  • Upload domain-specific terminology
  • Boost recognition accuracy for specialized content
  • Support for up to 10,000 custom terms per account

​

March 2025

​

v1.4.2 - March 25, 2025

• NEW: Audio preview before transcription
  • 30-second preview for uploaded files
  • Waveform visualization
  • Basic audio quality metrics
• IMPROVED: Error reporting
  • More detailed error messages
  • Suggested fixes for common issues
  • Error code documentation

​

v1.4.0 - March 12, 2025

• NEW: Advanced audio preprocessing
  • Automatic noise reduction
  • Volume normalization
  • Audio format conversion (automatic)
• NEW: Audio quality scoring
  • Quality score (1-10) for uploaded audio
  • Recommendations for improving transcription accuracy
  • Warning for low-quality audio files

• IMPROVED: Word-level timestamp accuracy
  • More precise timing information
  • Better alignment with audio
  • Support for subtitle formats (SRT, VTT)

​

February 2025

​

v1.3.5 - February 28, 2025

• NEW: Detailed usage breakdowns
  • Per-model usage statistics
  • Cost analysis by feature (diarization, summarization, etc.)
  • Monthly usage reports via email
• NEW: Usage alerts and limits
  • Set spending limits per month/day
  • Email/webhook notifications for usage thresholds
  • Automatic transcription pausing at limits

​

v1.3.0 - February 14, 2025

• IMPROVED: 30% better speaker identification accuracy
• NEW: Speaker labeling (Speaker A, Speaker B, etc.)
• NEW: Confidence scores for speaker assignments
• NEW: Support for up to 10 speakers (previously 6)

• NEW: Custom summary prompts
  • Define your own summarization style
  • Industry-specific summary templates
  • Variable summary lengths
• IMPROVED: Summary quality across all languages
• NEW: Key points extraction in addition to summaries

​

January 2025

​

v1.2.8 - January 30, 2025

• NEW: Multi-region deployment
  • Reduced latency for global users
  • Data residency options for EU customers
  • Automatic failover between regions
• IMPROVED: 99.95% uptime SLA (up from 99.9%)
• IMPROVED: Processing capacity increased by 3x

​

v1.2.0 - January 15, 2025

• IMPROVED: 50% faster API response times
• IMPROVED: Better handling of concurrent requests
• NEW: Health check endpoints for monitoring
• FIXED: Memory leaks in long-running transcriptions

​

v1.1.0 - January 5, 2025

• NEW: GDPR compliance features
  • Data processing agreements
  • Right to erasure (delete all user data)
  • Data portability (export user data)
• NEW: SOC 2 Type II compliance
• NEW: End-to-end encryption for file uploads
• IMPROVED: Enhanced PII detection and redaction

​

Migration Notes

​

From v1.4.x to v1.5.x

# Old format (deprecated)
curl -H "Authorization: sk-your-api-key" https://api.verbalisai.com/v1/transcript

# New format (required)
curl -H "Authorization: Bearer sk-your-api-key" https://api.verbalisai.com/v1/transcript

// Old format
{
  "transcription_id": "12345",
  "status": "completed"
}

// New format
{
  "event": "transcription.completed",
  "data": {
    "transcription_id": "12345",
    "status": "completed"
  },
  "timestamp": "2025-04-16T10:30:00Z",
  "version": "1.5.0"
}

​

Upcoming Features

• Real-time streaming transcription (Q3 2025)
• Custom model training (Q4 2025)
• Multi-modal analysis (audio + video) (Q4 2025)
• Advanced analytics API (Q3 2025)