Open-source audio transcoding proxy that renders variants on demand from signed URLs
AI-powered voice call investigation and root cause analysis for Slack
Stage
Not given
Startup
Founded
Not given
2025
Based in
Not given
ES
Pricing model
Free
Freemium
Pricing
Open source and free to self-host
Starter: $0/month with 100 free credits (one-time) per Slack workspace. Team: $50/month for 50,000 credits (auto-refreshing monthly), scaling to $5,000/month for 5,000,000 credits. Enterprise: custom pricing available on request.
Key features
On-demand audio transcoding from signed URLs
Multiple output formats: MP3, AAC, M4A, Opus, Ogg, FLAC, WAV, and peaks
Audio trimming, fades, gain and EBU R128 loudness normalization
Speech cleanup with enhance:voice preset
Waveform peaks generation compatible with audiowaveform
Expiring URLs excluded from cache key
Render stream while encoding with real-time caching
Range-capable seeking for cached variants
Immutable cache with long-lived variant URLs
Audio-only enforcement preventing video processing
Natural language questions answered in under 5 seconds
Cross-service correlation across 20+ voice and operations providers
Transcript search and conversation analysis
Call timeline correlation with ad campaigns and CRM data
OAuth-only integration requiring no code changes or SDK installation
HubSpot CRM integration for caller enrichment
What makes it different
URL-based API with no request bodies or server-side state
Options normalize to same cache key regardless of order
Streams first render to client while caching simultaneously
Range-seeking on cached variants with CDN-friendly immutable policies
Open source with Elixir implementation and reference signing algorithms
Answers posted in Slack threads within 5 seconds without leaving the chat interface
Correlates data across multiple providers in one sourced verdict instead of tab-switching
Zero code changes, SDKs or custom configuration required
Free tier with no credit card requirement to get started
Built by operators for voice AI teams, not generic observability tool
audioproxy
Sherlock Calls