voice-mode
Activates voice conversation mode using Pocket TTS Docker container. Use when user says "voice mode", "let's talk", "talk to me", "speak your responses", or wants Claude to respond with spoken audio. Speaks all responses through TTS and plays via speakers.
Best use case
voice-mode is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Activates voice conversation mode using Pocket TTS Docker container. Use when user says "voice mode", "let's talk", "talk to me", "speak your responses", or wants Claude to respond with spoken audio. Speaks all responses through TTS and plays via speakers.
Teams using voice-mode should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/voice-mode/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How voice-mode Compares
| Feature / Agent | voice-mode | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Activates voice conversation mode using Pocket TTS Docker container. Use when user says "voice mode", "let's talk", "talk to me", "speak your responses", or wants Claude to respond with spoken audio. Speaks all responses through TTS and plays via speakers.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Voice Mode
Voice conversation mode where all responses are spoken aloud via Pocket TTS.
## Setup
The `tts.sh` script lives in this skill's `scripts/` directory. Resolve it relative to this SKILL.md:
```
SKILL_DIR="<absolute path to this skill's directory>"
TTS="${SKILL_DIR}/scripts/tts.sh"
```
Use `${TTS}` for all commands below.
## Activation
On activation, ALWAYS run these steps in order before anything else:
1. Check the TTS container is running:
```bash
${TTS} ensure
```
If this fails (exit code 1), tell the user the container is down and stop. Do NOT attempt to start it.
2. Confirm voice mode is active by speaking:
```bash
${TTS} play "Voice mode activated. I'm listening." -v eponine
```
## Response Rules
While voice mode is active:
1. **ALWAYS speak every response** using tts.sh:
```bash
${TTS} play "<response text>" -v eponine
```
2. **Prefer concise responses** — aim for 1-3 sentences when used standalone. When combined with another skill, match the response length that skill requires.
3. **Write naturally for speech** — avoid markdown, bullet points, code blocks, URLs. Write as you'd speak in conversation.
4. **Also output text** — print a brief text version so the conversation is readable in the terminal.
5. **Handle STT input gracefully** — user input arrives as `[STT]...[/STT]` tags from their whisper script. The transcription may be imperfect. Infer intent from context rather than asking for clarification on every garbled word.
6. **Split long responses** — if you need to say more than ~2 sentences, make multiple tts.sh calls so audio starts playing sooner.
## Voice Selection
Default voice: `eponine`
If the user provided an argument (e.g., `/voice-mode jean`), use that voice instead.
Available: alba, marius, javert, jean, fantine, cosette, eponine, azelma
## Deactivation
Voice mode ends when the user says "stop voice mode", "text mode", or "stop talking".
Confirm with a final spoken message: "Voice mode off. Back to text."
## Configuration
All configurable via environment variables:
- `POCKET_TTS_PORT` — server port (default: 18731)
- `POCKET_TTS_VOICE` — default voice (default: eponine)
- `POCKET_TTS_SPEED` — playback speed (default: 1.2)
## Dependencies
- Docker with pocket-tts container running (`docker compose up -d` from the pocket-tts repo)
- mpv (audio playback)
- curlRelated Skills
workflow-review
Reviews Claude Code sessions and proposes workflow improvements. Use when: (1) /workflow-review command, (2) "review my workflow", "how can I improve", (3) after long sessions when nudged, (4) start of session with pending review. Analyzes tool usage patterns, CLAUDE.md configuration, and compares against CC best practices. Proposes: CLAUDE.md updates, new skills, underused CC features. Saves session summaries to .claude/workflow-reviews/ for cross-session continuity.
test-quality
Guides strong, effective unit test generation using proven testing techniques. Use when writing unit tests, reviewing test quality, improving existing tests, generating test cases, checking test coverage strength, or when tests exist but may be weak. Triggers on: unit test, test quality, test coverage, write tests, improve tests, review tests, test strength, mutation testing, boundary testing.
pr-review
Reviews code changes before merging. Use when reviewing PRs, checking staged changes, reviewing diffs, code review, merge readiness check, or validating changes before commit/push.
latex-presentation
Creates impressive LaTeX Beamer presentations with modern design. Generates .tex files with theme selection, font pairing, TikZ diagrams, overlays, and best practices. Use for LaTeX slides, Beamer presentations, scaffolding decks, slide design advice, or TikZ diagram generation.
git-commit
Plans and executes git commits with optional TICKET_ID prefix. Analyzes staged changes, proposes optimal commit structure (single or multiple), generates descriptive messages with technical context, and executes after user approval. Use when committing code changes, creating atomic commits, or splitting large changesets.
ctask
Manages tasks using the ctask CLI wrapper over a local SQLite database. Use when tracking work items, creating tasks, managing dependencies, adding comments, labeling, or reviewing task status. Triggers on task tracking, ticket management, work planning, backlog management.
codex
AI peer review via OpenAI Codex CLI. Use when reviewing code changes, validating technical decisions, comparing implementation approaches, or getting a second opinion on architecture choices. Triggers on /codex, /codex-review, or auto-triggers when presenting significant alternatives to user.
codemap
Generate navigational codebase maps with architecture diagrams. Use when mapping a codebase, creating architecture docs, visualizing project structure, generating infrastructure diagrams, understanding repo layout, or onboarding to a new project.
clarice
Conducts realistic mock interviews with detailed feedback and scoring. Use for interview prep, behavioral questions, technical interviews, STAR practice, system design interviews, or interview coaching.
c7
Fetches up-to-date library documentation from Context7 and saves to /tmp/context7/. Use when needing current API docs, code examples, library references, SDK documentation, or checking latest library versions. Triggers: context7, c7, library docs, fetch docs, current documentation, api reference.
statsmodels
Advanced statistical modeling and hypothesis testing. Complementary to SciPy's stats module, it provides classes and functions for the estimation of many different statistical models, as well as for conducting statistical tests and statistical data exploration. Use for linear regression, GLM, time series analysis, ANOVA, survival analysis, causal inference, and statistical hypothesis testing. Load when working with OLS, WLS, logistic regression, Poisson regression, ARIMA, SARIMAX, statistical diagnostics, p-values, confidence intervals, or R-style statistical analysis.
design-mode
Complement to frontend-design — embodies Anthropic's design-surface system prompt (the "expert designer" persona that produces HTML artifacts, decks, prototypes, and animations in a filesystem-based project). Use when the user wants to enter "design mode" and operate as a design-artifact generator with the full Anthropic design workflow (asking questions first, exploring design systems, Tweaks, speaker notes, starter components, verifier agent), or when frontend-design has set aesthetic direction and the next step is producing the artifact. Trigger phrases: "design mode", "design artifact", "make a deck", "hi-fi prototype", "Anthropic design mode", "Claude design sys prompt", "design like claude.ai/design".