SDKs
Install and configure the official Addis AI Node.js and Python SDKs.
New
The official addisai SDKs are the recommended way to call Addis AI from Node.js and Python backends.
Install
npm install https://github.com/Addis-AI-Org/addisai-js/releases/download/v0.4.0/addisai-0.4.0.tgzpip install "addisai[realtime]==0.4.0"Create a client
Set your key in the server environment:
export ADDIS_API_KEY="your_api_key"import AddisAI from "addisai";
const addis = new AddisAI();from addisai import AddisAI
addis = AddisAI()The SDKs read ADDIS_API_KEY by default. Keep this key on a trusted server—never in browser JavaScript, a mobile binary, or a public environment variable.
SDK capability matrix
| Capability | Node.js and Python interface |
|---|---|
| Chat and text generation | addis.chat.completions.create |
| Multi-turn chat | addis.chat.completions.create with messages |
| Function calling | addis.chat.completions.create and SDK tool runners |
| Speech-to-text | addis.speech.transcribe |
| Addis Scribe file transcription (0.4.0 update) | addis.scribe.transcribe |
| Scribe partial transcript stream (0.4.0 update) | addis.scribe.stream |
| Scribe live PCM sessions (0.4.0 update) | addis.scribe.connect |
| Scribe usage and request recovery (0.4.0 update) | addis.scribe.usage, addis.scribe.recover |
| Scribe word timestamps and caption cues (0.5.0 update) | addis.scribe.transcribe with timestamps: "word" |
| SRT and VTT captions from a Scribe result (0.5.0 update) | toSrt / toVtt (Node.js), addisai.to_srt / addisai.to_vtt (Python) |
| Scribe speaker labels, Turbo only (0.6.0 update) | addis.scribe.transcribe with backend: "turbo" and speakers: true |
| SRT and VTT captions with speaker labels (0.6.0 update) | toSrt / toVtt (Node.js), addisai.to_srt / addisai.to_vtt (Python) |
| Translation | addis.translate.create |
| Image and document reasoning | addis.chat.completions.create with attachments |
| Audio reasoning | addis.chat.completions.create with audio input |
| List voices | addis.voices.list |
| Preview a voice | addis.voices.preview |
| Estimate voice cost | addis.voice.estimate |
| Generate a voice clip | addis.voice.generate |
| Stream voice audio (0.3.0 update) | addis.voice.stream |
| Voice WebSocket sessions (0.3.0 update) | addis.realtime.connect |
| Check voice usage | addis.voice.usage |
| List and manage clips | addis.voice.clips |
Addis Voices 2 supports completed clips and Streaming text-to-speech. Microphone conversations use the separate Realtime API.
Choose an integration guide
The original integration guides remain the source of truth for application architecture: