What Tablif read on Cekura Bench's site
Cekura Bench describes itself as: “Cekura Bench”
- Stage: Live, no waitlist
What it is
Calls itself “voice AI benchmarks”.
- Calls itself: voice AI benchmarks
“…voice AI benchmarks”
Where it puts itself
Publishes a comparison page.
“Compare -> https://benchmarks.cekura.ai/speech-to-speech/compare/cascade-baseline-vs-openai-realtime”
What it costs
No monthly plan. 7 prices listed on its pages, from $0 to $0.
- Monthly: no
- Price: 7 prices, from $0 to $0
“Cost / min $0.020”
“Cost / min $0.055”
How it runs
Distributed through the GitHub. No API. Open source. AI is the core of the product.
- Distributed via: GitHub
- API: no
- Open source: yes
- AI role: core to the product
“Test cases, agent definitions and benchmark code are public on GitHub”
“The data behind the speech benchmarks can be downloaded as JSON.”
What it works with
Names Amazon Nova 2 Sonic, Azure Realtime and 5 more on its pages.
- Names: Amazon Nova 2 Sonic · Azure Realtime · Deepslate Opal · +4
“Amazon Nova 2 Sonic”
“Azure Realtime”
What it offers as proof
Live, no waitlist. Claims “9 realtime voice models”. Publishes docs and a blog.
- Number claimed: 9 realtime voice models
“9 realtime voice models”
“Documentation -> https://docs.cekura.ai/”
Sources
Tablif read these pages, last on 8 Oct 2026. Everything above is what the product says about itself; Tablif does not verify every claim.
Similar products in Tablif Town
- Maina Voice: Speech to text dictation and multi-engine benchmarking
- VoiceFlow AI: Free Ultra-Realistic AI Text to Speech Generator
- Orchard: Speech-to-text for LatAm calls. Most accuracy per dollar.
- Speechka: Real-time voice translation that sounds like you
- Sovereign Voice: On-device voice synthesis. No cloud, no subscription.
- Vozon: Ai Voice agent platform
- MAR8 - Text to Speech AI by CAMB.AI: Most AI voice systems look the same on the surface: a single Text-to-Speech model hidden behind an API.
- Cue v5: Help voice agents know when to stop, pause and respond