Speko
AIA router for voice AI that benchmarks 56 speech and language models language by language, then routes each session to the one that wins.
Speko
A router for voice AI that benchmarks 56 speech and language models language by language, then routes each session to the one that wins.
About Speko
Speko exists because voice models are almost always benchmarked in English while real users are not speaking English. It measures 56 speech and language models across ten languages — English, Arabic, French, German, Hindi, Norwegian, Spanish, Tamil, Telugu and more — and publishes the results as an open comparison of word error rate against cost per minute for each stage of a voice pipeline. The findings are the product's argument: eleven of twenty-three speech-to-text models are only measured in English at all, and four different models win across nine languages, so no single vendor is the right default. On top of that measurement layer sits a hosted, provider-neutral STT, LLM and TTS data plane at router.speko.dev with typed contracts, OpenAPI and AsyncAPI specs, and managed routing that selects a model per session and fails over before a response is returned. An open customer-side runtime called Gateway gives LiveKit and Pipecat native integrations with provider-direct streaming and local BYOK credentials, so teams can keep their own vendor relationships and still get measured routing, and an MCP server lets Claude or Cursor query the benchmarks directly. Speko is backed by Y Combinator.
Screenshots
Key Features
Use Cases
Frequently asked questions about Speko
What is Speko?
A router for voice AI that benchmarks 56 speech and language models language by language, then routes each session to the one that wins.
How does Speko work?
Speko works by combining Language-by-Language Benchmarks: 56 speech and language models measured across ten languages, exposing that a model topping the English table may have no published rank anywhere else., Cost-Versus-Accuracy Charts: Word error rate plotted against dollars per minute for each pipeline stage — STT, LLM, TTS and speech-to-speech — so a model choice is a priced trade-off, not a guess., Managed Router: A hosted, provider-neutral data plane at router.speko.dev that selects the winning model per session using the measurements and fails over before a response is returned., Typed Contracts: Public OpenAPI and AsyncAPI specifications for the Router, so integrations are generated rather than hand-rolled against an undocumented endpoint., Open Gateway Runtime: A customer-side runtime with native LiveKit and Pipecat services offering provider-direct streaming and local BYOK credentials, keeping your vendor relationships intact. to help users with Multilingual Voice Products: Ship a voice agent to non-English markets and route each language to the model that actually performs best there., Vendor Selection Research: Use the published benchmarks to decide which STT or TTS provider to sign with before committing to a contract., Voice Cost Optimization: Trade a small accuracy delta for a large cost reduction by picking a model at the right point on the WER-versus-price curve., Failover and Reliability: Keep voice sessions alive through a provider outage with pre-response failover across the routed model pool., LiveKit or Pipecat Migration: Add measured routing to an existing agent stack by swapping in Gateway services rather than rewriting the pipeline..
What are the main features of Speko?
Key features include Language-by-Language Benchmarks: 56 speech and language models measured across ten languages, exposing that a model topping the English table may have no published rank anywhere else., Cost-Versus-Accuracy Charts: Word error rate plotted against dollars per minute for each pipeline stage — STT, LLM, TTS and speech-to-speech — so a model choice is a priced trade-off, not a guess., Managed Router: A hosted, provider-neutral data plane at router.speko.dev that selects the winning model per session using the measurements and fails over before a response is returned., Typed Contracts: Public OpenAPI and AsyncAPI specifications for the Router, so integrations are generated rather than hand-rolled against an undocumented endpoint., Open Gateway Runtime: A customer-side runtime with native LiveKit and Pipecat services offering provider-direct streaming and local BYOK credentials, keeping your vendor relationships intact..
Who is Speko for?
Speko is useful for anyone interested in Multilingual Voice Products: Ship a voice agent to non-English markets and route each language to the model that actually performs best there., Vendor Selection Research: Use the published benchmarks to decide which STT or TTS provider to sign with before committing to a contract., Voice Cost Optimization: Trade a small accuracy delta for a large cost reduction by picking a model at the right point on the WER-versus-price curve., Failover and Reliability: Keep voice sessions alive through a provider outage with pre-response failover across the routed model pool., LiveKit or Pipecat Migration: Add measured routing to an existing agent stack by swapping in Gateway services rather than rewriting the pipeline..
How much does Speko cost?
Speko offers a free tier with paid plans for advanced features.
How do I get started with Speko?
Visit https://speko.ai to sign up and explore Speko.
Explore more AI Ai Services tools
Browse all Ai Services tools →Compare Speko: vs oMLX · vs Olostep · vs FreeLLMAPI · vs PageIndex