A free skill for your next project

Speech and Voiceover

Generate narrated audio from approved text with a reusable speech CLI.

What it does

Generate narrated audio from approved text with a reusable speech CLI. Python, the OpenAI SDK, OPENAI_API_KEY and billable Audio API access. A local install alone does not generate audio. Review generated media and local command behavior before sharing or publishing.

Best for

  • Create a calm voiceover from the exact supplied 30-second product-demo script.

Output examples

No recorded run is published for this skill yet. The example below is illustrative, not a measured result.

Example input

Create a calm voiceover from the exact supplied 30-second product-demo script.

Illustrative output

Illustrative: A speech audio file with the original text and voice settings retained.

Requirements & permissions

  • Python, the OpenAI SDK, OPENAI_API_KEY and billable Audio API access. A local install alone does not generate audio.
  • Upstream examples use the legacy ~/.codex/skills location. Set SPEECH_CLI / TRANSCRIBE_CLI to the actual script in your chosen agent install folder before running example commands. The upstream repository is deprecated; evaluate updates independently.

What it may access

  • Read user-authorized source text/audio/video
  • Write generated files
  • Run inspected local CLI commands
  • Network access to external services; user-supplied credentials stay local
  • Audio/text sent to the OpenAI API for live jobs; API charges are separate.

The package is free. Your agent, model subscription, external API usage and local software may have separate costs. Review scripts before running them, and use copies of important files.

How to install

Download the ZIP, extract the skill folder, and follow the instructions for your agent. Keep the complete folder together.

Codex

Supported by upstream documentation; not independently tested in this agent. No end-to-end install test. Upstream skill format and documentation reviewed; tool names and legacy script paths may require adaptation.

  1. Download speech-1.0.0.zip and verify its SHA-256.
  2. Extract the complete speech folder into .agents/skills/speech/. Keep its license and supporting files.
  3. Confirm .agents/skills/speech/SKILL.md exists. Reload your agent skill list if needed.
  4. Read the requirements and review commands. Use the scripts from this actual installed directory; credentials and paid service setup are separate. Ask the agent to use speech.

Read the full Codex installation guide →

Claude Code

Not verified in this agent. Treat these as the agent’s standard manual installation steps, not a compatibility guarantee. No end-to-end install test. Upstream skill format and documentation reviewed; tool names and legacy script paths may require adaptation.

  1. Download speech-1.0.0.zip and verify its SHA-256.
  2. Extract the complete speech folder into .claude/skills/speech/. Keep its license and supporting files.
  3. Confirm .claude/skills/speech/SKILL.md exists. Reload your agent skill list if needed.
  4. Read the requirements and review commands. Use the scripts from this actual installed directory; credentials and paid service setup are separate. Ask the agent to use speech.

Read the full Claude Code installation guide →

Cursor

Not verified in this agent. Treat these as the agent’s standard manual installation steps, not a compatibility guarantee. No end-to-end install test. Upstream skill format and documentation reviewed; tool names and legacy script paths may require adaptation.

  1. Download speech-1.0.0.zip and verify its SHA-256.
  2. Extract the complete speech folder into .cursor/skills/speech/. Keep its license and supporting files.
  3. Confirm .cursor/skills/speech/SKILL.md exists. Reload your agent skill list if needed.
  4. Read the requirements and review commands. Use the scripts from this actual installed directory; credentials and paid service setup are separate. Ask the agent to use speech.

Read the full Cursor installation guide →

Source & license

Created by OpenAI. This third-party package is curated and redistributed under Apache-2.0. SkillPick is an independent distributor and is not endorsed by the original author.

Keep the package’s license and required notices with any copies. Read the bundled license and notices.

Review & limitations

Source and redistribution review only, not a security certification. Complete per-directory Apache-2.0 license retained. Script source inspected for API calls, file writes, environment key usage and dependencies. No live API call, audio generation or transcription was performed. Deprecated upstream status and legacy CLI paths are disclosed.

Last source review: . Package release: 1.0.0.

Keep exploring

Related skills for the work ahead.

All skills ↗