stt-stream

A command-line tool for turning an audio file into text as it arrives over a WebSocket connection, using xAI's streaming speech-to-text service.

In plain words
What is it for?
Transcribing WAV or raw PCM audio files, choosing a language, printing JSON events, and testing streaming transcription settings such as sample rate and audio encoding.
Why use it?
It lets scripts receive transcription events while audio is being processed instead of waiting for one final result. It also supports temporary, in-progress results when requested.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/moore-developers/grok-cli/stt-stream
Clone the repo
git clone --depth 1 https://github.com/Moore-developers/grok-cli
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 652 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00652
Opus 5 $0.00000 $0.00326
Sonnet 5 $0.00000 $0.00130
Haiku 4.5 $0.00000 $0.00065

Measured 2d ago against content hash 886722723d1f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

stt-stream scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/commands/stt-stream.md · 75 lines

How it starts

The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.

grok-cli stt-stream

Purpose

Run real-time transcription through the xAI WebSocket STT interface.

This is an experimental entrypoint and does not replace the batch transcription command stt. The first version sends a local audio file as binary frames over WebSocket, then sends {"type":"audio.done"}, and keeps reading transcription events until the stream ends.

Common Usage

grok-cli stt-stream --file ./sample.wav --language en --interim-results

Print JSON event summaries:

grok-cli stt-stream --file ./sample.wav --json

Use raw PCM parameters:

grok-cli stt-stream --file ./sample.raw --encoding pcm_s16le --sample-rate 16000 --language en

Parameters

  • PATH: positional audio file to transcribe.
  • --file <PATH>: explicit script-friendly audio file input.
  • --json: use the standard JSON envelope. data.events contains the received event list.
  • --auth-file <PATH>: override the OAuth state file path.
  • --model <MODEL>: override the streaming STT model, default grok-transcribe.
  • --language <LANG>: language code, default en.
  • --interim-results: request interim transcription results.
  • --endpointing <VALUE>: pass through the official endpointing parameter.
  • --encoding <ENCODING>: raw audio encoding, such as pcm_s16le.
  • --sample-rate <HZ>: raw audio sample rate.
  • --diarize: enable speaker diarization.
  • --filler-words: keep filler words.
  • --multichannel: treat audio as multichannel.
  • --channels <LIST>: choose channels such as 0,1.
  • --keyterm <TERM>: keyword boosting. Repeatable.
  • --timeout <SECONDS>: reserved WebSocket session timeout parameter. It is mainly retained for CLI compatibility.

Behavior

  • The real connection URL is derived from the OAuth state base_url, for example https://api.x.ai/v1 becomes wss://api.x.ai/v1/stt.
  • Streaming STT configuration parameters go into the URL query string.
  • Access token expiry is checked before the request. If needed, the command refreshes first.
  • Non-JSON output prints interim: ... or final: ... events.
  • JSON output returns the event array after the connection ends. Each event preserves normalized fields and the raw event.
  • This is streaming STT; it does not replace the batch stt multipart file / URL transcription path.

Read the full file on GitHub · 75 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 75 lines · 0 tokens per session scan A 886722723d1f

Subscribe to this mod's changes

stt-stream is a command published in the GitHub repository Moore-developers/grok-cli (52 stars, last pushed 19d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 652 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.