Build with the API
Call the API directly from a terminal, a service, or a notebook. Best when you own the code and want full control.CLI
voiceai runs speech from a terminal and manages agents and other resources. Good for scripts and CI.JavaScript SDK
voiceai-sdk on npm. A typed client for Node, Bun, and Deno.Python SDK
voiceai-sdk on PyPI. Sync and async clients for Python 3.9+.Hand it to a coding agent
Let a coding agent like Claude Code write the integration for you, or migrate an existing project onto SLNG.Agent skills
Teach a coding agent to call the API: speech, transcription, and voice agents.
Migration skills
Move a LiveKit, Pipecat, or custom agent onto SLNG hosted models.
Add to an existing agent
Already running a voice agent on a framework? Swap its speech models for SLNG and keep the rest of your pipeline.
LiveKit
The
livekit-plugins-slng adapter for LiveKit Agents.
Pipecat
The
pipecat-slng STT and TTS services for Pipecat pipelines.
n8n
The
n8n-nodes-slng community node for visual automations.Your SLNG key
Every tool reads its key from an environment variable. Create one at app.slng.ai, or see Create your API key, then export it before running the examples.The CLI reads
VOICEAI_API_KEY. The JavaScript and Python SDKs default to
SLNG_API_KEY but accept an explicit apiKey option. Set both to the same
value if you plan to use several tools side by side.Model identifiers
Every tool takes the same model identifier format,provider/family:version-lang.
Prefix with slng/ to target an SLNG-hosted, in-region model
(slng/deepgram/nova:3-en); a bare provider/model is the external or BYOK
route proxied through SLNG on your own provider_key. Browse the current models
in the catalog.