> ## Documentation Index
> Fetch the complete documentation index at: https://docs.slng.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent skills

> Install the slng-ai/skills pack so coding agents like Claude Code can call text-to-speech, transcription, and voice-agent APIs on your behalf.

`slng-ai/skills` is a bundle of agent skills for SLNG. After you install it, your coding agent knows how to set up an SLNG key, synthesize speech, transcribe audio, build voice-agent prompts, and dispatch outbound calls. Each skill is a short markdown brief plus example code the agent can run directly.

## Install

```bash theme={null}
npx skills add slng-ai/skills
```

That command fetches the latest version of every skill and drops them into the skills directory your agent reads from (e.g. `~/.claude/skills/` for Claude Code). Re-run it to update.

## Set up your key

The skills expect `VOICEAI_API_KEY` to be set in your shell or in a project `.env`:

```bash theme={null}
export VOICEAI_API_KEY="zpka_..."
```

Get one from [app.slng.ai](https://app.slng.ai/api-keys). The `setup-api-key` skill walks the agent through this on first use.

## What you get

| Skill            | What the agent learns to do                                                 |
| ---------------- | --------------------------------------------------------------------------- |
| `setup-api-key`  | Configure and validate `VOICEAI_API_KEY`                                    |
| `text-to-speech` | Generate speech with Deepgram Aura, Rime Arcana, Cartesia, Murf, and others |
| `speech-to-text` | Transcribe audio with Deepgram Nova, Sarvam Saaras, Soniox, or Reson8       |
| `agent-prompt`   | Build a complete voice-agent prompt ready to paste into the Agent Builder   |
| `agents`         | Create, manage, and dispatch SLNG voice agents (including outbound calls)   |

Each skill shows the same task four ways: CLI, JavaScript SDK, Python SDK, and raw HTTP. The agent picks whichever fits the project it is working in.

## Example: ask your agent to transcribe a file

After installing the pack, ask the agent something like:

> Transcribe `meeting.wav` with slng and summarize the result.

The agent picks up the `speech-to-text` skill, runs `voiceai stt meeting.wav` (or the SDK equivalent if the project is already Python or JavaScript), parses the response, and writes the summary back to you.

## Works with

Any coding agent that loads skills from a directory. The skills are plain markdown plus shell-runnable examples, so they work with Claude Code, Cursor's agent mode, and any compatible host.

## Source

* Repo: [github.com/slng-ai/skills](https://github.com/slng-ai/skills)
* File an issue or PR there to propose new skills.