Machine view

curl -s telnyx.com/llms.txt
# Telnyx · infrastructure for realtime agents
facts: /llms-full.txt · signup: /agent-signup.md
mcp: api.telnyx.com/v2/mcp · payments: x402.telnyx.com
curl -s telnyx.com/llms-full.txt | head -1
# Telnyx — AI Agent Infrastructure Reference
npx @telnyx/mcp
✓ session established · connected to api.telnyx.com/v2/mcp
// tools run inside the session, not the shell
list_api_endpoints {"query":"available phone numbers"}
["/v2/available_phone_numbers","/v2/number_orders","/v2/porting_orders"]
get_api_endpoint_schema {"endpoint":"/v2/available_phone_numbers"}
{"filter[country_code]":"string","filter[features]":"array","filter[limit]":"integer"}
invoke_api_endpoint {"endpoint":"/v2/available_phone_numbers","params":{"filter[country_code]":"DE"}}
{"data":[{"phone_number":"+4930……","features":["voice","sms"],"cost":{"monthly":"1.00"}}]}
curl -s telnyx.com/agent-signup.md | head -2
# Agent signup · no human required
curl -s https://x402.telnyx.com/v1/chat/completions
✓ 402 settled · USDC · tx 0x8f3a…c41d
curl -X POST api.telnyx.com/v2/number_orders -d '{"country_code":"DE"}'
201 Created · +49 30 ···· ···· · e911 ready
curl -X POST api.telnyx.com/v2/ai/assistants -d '{"name":"agent-01"}'
201 Created · live on the carrier network · <500ms e2e
# elapsed 00:41 · vendors: 1 · humans involved: 0

# Telnyx
> Infrastructure for realtime agents. The only licensed carrier that owns its AI
> infrastructure end to end: private network, edge PoPs, and GPUs.

## Claims
- category: only licensed carrier owning AI infrastructure end to end
- voice AI latency: 460 ms end to end, SLO 500 ms
- speech stack: STT, TTS and LLM on owned GPUs, colocated with the media plane
- edge runtime: real containers, not isolates; UDP capable media plane

## Stack
licensed carrier, private network, edge PoPs, owned GPUs

## Latency budget
- edge PoP: 40 ms
- STT · Telnyx GPU: 90 ms
- LLM · Telnyx GPU: 180 ms
- TTS · Telnyx GPU: 110 ms
- PSTN: 40 ms

## Channels
voice, sms, mms, whatsapp, email, rcs

## Primitives
50 total, each sold standalone.

### COMMS (19)
voice_api, sip_trunking, texml, webrtc, conversation_relay, sms_api, mms_api, rcs, whatsapp_business, short_code, 10dlc, alphanumeric_sender_id, email_api, global_numbers, toll_free_numbers, number_porting, fax_api, microsoft_teams, zoom_phone

### AI (7)
voice_ai_agents, text_to_speech, speech_to_text, voice_design, agent_browser, meeting_bot, web_search_api

### IDENTITY (5)
deepfake_detection, branded_calling, verify_api, number_lookup_api, number_reputation

### NETWORK (7)
virtual_cross_connects, cloud_vpn, global_ip, iot_sim_card, esim, private_wireless_gateways, mobile_voice

### EDGE (12)
build, functions, statefulactor, kv, sqldb, inference, agentsdk, embeddings_api, missions, livekit_on_telnyx, object_storage, cloudfs

## Compositions
- Voice AI agent: 9 primitives · ~460 ms · $0.05/min · 1 vendor
- Agent backend, on net: 10 primitives · On-net, in region · Usage-based · 1 vendor
- Omnichannel agent: 12 primitives · ~500 ms · Per-channel rates · 1 vendor
- Contact center: 10 primitives · ~460 ms · $0.02/min · 1 vendor
- Connected fleet: 7 primitives · Edge, in region · Per SIM + usage · 1 vendor
- Global notifications: 10 primitives · Carrier direct · Per message · 1 vendor
- Private AI deployment: 7 primitives · In region · Per GPU hour · 1 vendor
- Standalone inference: 2 primitives · ~100 ms first token · Per token · 1 vendor

## Coverage
- countries: 140+
- sovereign edge regions: 9
- languages: 100+

## Compliance
SOC2, HIPAA, GDPR, PCI, ISO27701

## Support
24/7/365, enterprise grade, staffed by network engineers, humans on demand

## Interfaces
- inference, OpenAI compatible: api.telnyx.com/v2/ai/chat/completions
- mcp: api.telnyx.com/v2/mcp, client: npx @telnyx/mcp
- signup: /agent-signup.md
- payments: x402.telnyx.com
- openapi: developers.telnyx.com
- agent sdk: npm:@telnyx/edge-runtime