Overview

Convert LLM streaming responses into T.140 real-time text for SIP, WebRTC, and (S)RTP.

t140llm takes a streaming response from any LLM SDK and emits it as ITU-T T.140 real-time text — character by character, as the tokens arrive — over WebSocket, RTP, SRTP, or a Unix socket. It's the bridge between a modern chat completion stream and the real-time text infrastructure used by SIP, WebRTC, TTY/telecommunications relay, and assistive devices.

Install

npm install t140llm

Requires Node.js >= 16.

Quick start

import { processAIStream } from "t140llm";
import { OpenAI } from "openai";

const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

const stream = await openai.chat.completions.create({
  model: "gpt-4",
  messages: [{ role: "user", content: "Write a short story." }],
  stream: true,
});

processAIStream(stream); // T.140 over WebSocket (ws://localhost:8765)

Or send straight to an RTP endpoint:

import { processAIStreamToRtp } from "t140llm";

const transport = processAIStreamToRtp(stream, "192.168.1.100", 5004);
// transport.close() when done

Streams are consumed as-is — EventEmitters and async iterables both work, no wrapping required. The create-stream / hand-to-processAIStream* pattern is the whole surface; see the provider guide for every supported SDK.

Why T.140?

T.140 transmits text over IP as it is being written, rather than after the full message is composed — a good fit for low-latency and lossy links (accessibility services, TTY relays, noisy radio, satellite). The RTP pipeline uses an event-driven send model with sub-millisecond chunk-to-wire latency (avg ~0.1 ms in the included benchmark).

Transports

Function	Output	Notes
processAIStream	WebSocket	Default; T.140 over ws://
processAIStreamToRtp	RTP over UDP	Optional FEC (RFC 5109) and redundancy
processAIStreamToSrtp	Encrypted SRTP	Keys from passphrase or your own
processAIStreamToDirectSocket	Unix SEQPACKET socket	No WebSocket hop, still RTP-framed
processAIStreamsToMultiplexedRtp	One RTP stream, many LLMs	CSRC-tagged; demux on the far end

Every transport can be pre-connected before the stream exists (to cut startup latency) and can run over a custom transport (WebRTC data channel, steganographic carrier, etc.).

How it works

WebSocket mode — text chunks arrive over a WebSocket, get encapsulated as T.140 in RTP packets, and are written to a Unix SEQPACKET socket your application reads from.

Direct RTP mode — each chunk is packaged as T.140 in an RTP packet and sent over UDP to the target address. With FEC enabled, XOR-based FEC packets (RFC 5109) are generated per group so receivers can recover from loss.

Secure SRTP mode — the same RTP flow, encrypted with SRTP using the keys you supply.

Next steps

• Provider guide — OpenAI, Anthropic, Gemini, Mistral, Cohere, Ollama, Vercel AI SDK
• API reference — all functions, classes, and config
• Types & interfaces — RtpConfig, SrtpConfig, TransportStream, errors, metadata
• Steganography — hide RTP packets in cover media
• Examples — runnable demos