Hacker News Reader

fionaattelnyxon Run your agent's code where the LLM, TTS, and calls run

Today we shipped Telnyx Edge Compute, bringing the agent runtime to the same network as our inference, speech, and telephony. With everything agents need on one platform, agent latency can be even further reduced. Here's what we've launched:

- Functions (GA): containerized edge compute, deployed with telnyx-edge ship. Reaches other platform components through bindings, no API keys, auth headers, or network calls between them. - KV (GA): globally distributed key-value storage, accessed through bindings with no egress fees. - SQLDB: SQLite databases you create, bind in your config, and query from your functions. No database server to manage. - StatefulActor (beta): durable per-entity state with single-threaded execution. One instance per entity, one call at a time. - CloudFS (beta): shared POSIX filesystem backed by Cloud Storage. Mount on any container, files persist across restarts.

These join our Cloud STorage and Inference products which have been live on the platform for a number of years.

The link goes to a new microsite we built which itself is an Edge Compute Function, and allows you to run live demos in your browser, no sign up required.

You can also check out the docs: https://developers.telnyx.com/docs/edge-compute/platform-ove...

Please post feedback and what you'd like to see next.

fionaattelnyxon Kimi K3 Now Available via Telnyx Inference API

We'll be pushing benchmarks over the next few days, will update here when we have them

fionaattelnyxon Kimi K3 Now Available via Telnyx Inference API

Yes, we offer BAAs for workloads on our platform

fionaattelnyxon [dead]

We just launched GPU-powered Points of Presence in Paris, bringing sub-200ms round-trip times for conversational AI to European users.

  Why this matters for Europe:
  • **Latency-critical use cases**: Customer service, telehealth, real-time translation - all need <200ms for natural conversation
  flow
  • **GDPR compliance**: Data processed within EU borders, full regulatory compliance
  • **Local language optimization**: Better performance for European languages and accents
  • **Business continuity**: No more routing voice AI through US servers

  Technical details:
  • Sub-200ms round-trip time from major EU cities
  • GPU infrastructure optimized for real-time voice processing
  • Pricing: $0.05/minute for conversational AI
 
 Would love feedback from European developers, what voice AI use cases have you been waiting to build?