speech-setup

A setup guide for Claude Speech, a plugin that lets Claude speak responses aloud. It covers installing dependencies, configuring a Deepgram voice service or macOS's built-in speech, choosing a voice mode, and testing the setup.

In plain words
What is it for?
Use it when installing or changing Claude Speech, setting its Deepgram API key, selecting a voice, regenerating its hook configuration, or running a smoke test.
Why use it?
It turns the speech plugin's configuration into a guided sequence and checks whether the required software and settings are available.

Skill for Claude CodeCodex

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 skills/jason-c-dev/claude-code-speak/speech-setup
Any agent
npx skills add jason-c-dev/claude-code-speak --skill speech-setup
Clone the repo
git clone --depth 1 https://github.com/jason-c-dev/claude-code-speak

Made for: Claude Code, Codex.

Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,448 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 3 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 $0.00038 $0.01448
Opus 5 $0.00019 $0.00724
Sonnet 5 $0.00008 $0.00290
Haiku 4.5 $0.00004 $0.00145

Measured yesterday against content hash 9b3f06d006f7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

speech-setup scanned grade C with 3 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 yesterday.

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.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

chmod 600 ~/.claude/voice/.env

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

test -s ~/.claude/voice/.env && grep -q '^DEEPGRAM_API_KEY=' ~/.claude/voice/.env && echo HAVE || echo MISSING

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

import subprocess; subprocess.run(['afplay', str(p)])
skills/speech-setup/SKILL.md · 173 lines

How it starts

The opening of the file, as written. The whole thing — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Claude Speech — Setup Skill

You are walking the user through configuring (or reconfiguring) the Claude Speech plugin. Be concise. One short prompt at a time. Use the Bash tool for commands and the Edit/Write tools for files. Do NOT speak to them via the plugin during setup — that's the smoke test at the end.

The plugin lives at ${CLAUDE_PLUGIN_ROOT}. Per-user data and secrets live at ~/.claude/voice/.

Step 1 — Pre-flight

Run:

python3 --version

Confirm Python ≥ 3.10. If lower, stop and tell the user to install a newer Python.

Then:

python3 -c "import claude_agent_sdk" 2>/dev/null && echo OK || python3 -m pip install claude-agent-sdk

If it fails, fall back to pip install claude-agent-sdk and report the error if that also fails.

Step 2 — Deepgram API key (optional)

Check whether ~/.claude/voice/.env already contains DEEPGRAM_API_KEY. Use:

test -s ~/.claude/voice/.env && grep -q '^DEEPGRAM_API_KEY=' ~/.claude/voice/.env && echo HAVE || echo MISSING

If MISSING, ask the user:

"Do you have a Deepgram API key? (Paste it now to enable Aura-2 voice quality, or say 'skip' to use the macOS say fallback only.)"

If they paste a key, write it to ~/.claude/voice/.env:

mkdir -p ~/.claude/voice
printf 'DEEPGRAM_API_KEY=%s\n' "<KEY_FROM_USER>" >> ~/.claude/voice/.env
chmod 600 ~/.claude/voice/.env

If they skip, set primary_tts: "say" in the next step's config.

Step 3 — Voice picker

Present these six curated Aura-2 voices with one-liners:

  • aura-2-thalia-en — clear, confident, energetic American female (default)
  • aura-2-orion-en — approachable American male
  • aura-2-luna-en — friendly young-adult American female
  • aura-2-zeus-en — deep, trustworthy American male
  • aura-2-pandora-en — smooth, calm British female
  • aura-2-asteria-en — knowledgeable, energetic American female

Ask: "Which voice would you like? (Or paste any other Aura-2 model id, e.g. aura-2-callista-en.)" Default to aura-2-thalia-en.

Read the full file on GitHub · 173 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. yesterday First seen · 173 lines · 38 tokens per session scan C 9b3f06d006f7

Subscribe to this mod's changes

speech-setup is a skill published in the GitHub repository jason-c-dev/claude-code-speak (2 stars, last pushed 4mo ago), licensed MIT. It adds 38 tokens to every session and 1,448 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 3 findings (asks for root, reads agent configuration directories, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.