Direct competitors and differentiators
- Deepgram (Nova‑3/Flux): Known for sub‑300 ms streaming latency, accuracy/cost tradeoffs on real‑time, add‑ons (redaction, diarization), and per‑second billing; broad language coverage in streaming. Often used for production voice agents where price‑latency balance is important.
Source: Deepgram pricing page and comparisons.
- OpenAI (gpt‑4o‑transcribe / -mini‑transcribe; Whisper): Low cost ($0.003–$0.006/min); ecosystem hooks; growing diarization options. Trade‑offs include changing model availability and fewer enterprise‑grade logging controls vs ElevenLabs’ Zero Retention/data residency options.
Sources: OpenAI pricing/docs; HN pricing discussion.
- AssemblyAI (Universal‑Streaming): Lists $0.15/hr ($0.0025/min) with claims of very fast streaming; positions unlimited concurrency; offers self‑hosted/VPC deployments and HIPAA/BAA. Effective rates depend on session billing details.
Source: AssemblyAI pricing page.
- Hyperscalers (Google/AWS/Azure): Broad language/tooling ecosystems; predictable enterprise procurement; pricing varies (AWS often pricier at list for streaming).
Sources: Google/AWS/Azure pricing pages, Deepgram comparative article.
Feature comparison (high level)
- Strengths:
• Multilingual low‑latency (<150 ms) with reported accuracy; enterprise privacy options (Zero Retention), data residency (EU/India), HIPAA with BAA; VAD and manual commit features; supports PCM 8–48 kHz and μ‑law.
• Integration with ElevenLabs Agents (voice stack: STT + TTS + orchestrations).
- Where competitors lead:
• Price: OpenAI and AssemblyAI list lower per‑minute rates; Deepgram prerecord cheaper than Scribe v1.
• Deployment: AssemblyAI advertises self‑hosted/VPC; Deepgram offers enterprise SLAs and add‑ons.
• Open‑source/on‑device: Whisper ecosystem for offline/edge and cost control.
Pricing comparison (indicative list rates)
- Scribe v1 (batch): historically $0.40/hr (~$0.0067/min) in communications; plan‑based hours with overage.
- Deepgram Nova‑3 prerecord: ~$0.0043/min; streaming ~$0.0077/min.
- OpenAI gpt‑4o‑mini‑transcribe: ~$0.003/min; Whisper: ~$0.006/min.
- AssemblyAI Universal‑Streaming: ~$0.15/hr (~$0.0025/min) per site; terms/real‑world effective billing may vary.
Sources: vendor pricing pages.
Customer testimonials comparing products
- External roundups vary: some independent posts place OpenAI/Whisper or Gemini at/near top in specific conditions; others claim Scribe leads multilingual accuracy. Benchmarks are inconsistent across datasets; buyers should test on their audio.
Sources: Voice Writer leaderboard; independent blogs; ElevenLabs research post; press.
Market share and adoption trends
- ElevenLabs is a widely recognized brand in AI audio; ARR reportedly surpassed $200M by Nov 2025 and the tender valued the company at $6.6B. Adoption grew via publishers, creators, and enterprise pilots; Scribe extends that footprint into ASR.
Sources: ElevenLabs tender blog; Reuters/TechCrunch.
When customers choose competitors (and why)
- Price‑sensitive, very high‑volume batch workloads (OpenAI gpt‑4o‑mini‑transcribe or Deepgram prerecord).
- Strict on‑prem/VPC mandates (AssemblyAI self‑hosted; Speechmatics).
- Edge/offline use (Whisper on device).
- Required audio intelligence add‑ons priced à la carte with existing vendor.