hskhelper/aistackchan-hrs

Alternative implementation of AI Stack-chan

AIStackchan-hrs

Summary

This is an alternative implementation of AI Stack-chan. Super thanks to meganetaaan the originator of Stack-chan and robo8080 the originator of AI Stack-chan.

Features

• The following speech services can be used
  • Google Translate Text-to-Speech API (no API Key reqiured) - unofficial?
  • VoiceText Web API (API Key required) - free registration suspended for now
  • TTS QUEST V3 VOICEVOX API (API Key required)
• API
  • Speak API
  • Chat API (OpenAI API Key required)
• Speak randomly
  • send a question to the ChatGPT and speak answer periodically
  • start/stop by button A
• Speak clock
  • on every hour
  • current time by button C

Settings

Store the file named settings.json on SD card in advance (OPTIONAL). Once the stack-chan is running and connected to network, you can update settings via API.

curl -X POST "http://(Stack-chan's IP address)/settings" \
    -H "Content-Type: application/json" \
    -d @settings.json

You can also update some settings indivisually.

curl -X POST "http://(Stack-chan's IP address)/settings" \
    -d "voice.lang=en_US" \
    -d "voice.service=google-translate-tts"

Here is the example of settings.

{
  "network": {
    "wifi": {
      "ssid": "SSID",
      "pass": "PASSPHRASE"
    },
    "hostname": "stackchan"
  },
  "time": {
    "zone": "JST-9",
    "ntpServer": "ntp.nict.jp"
  },
  "servo": {
    "pin": {"x": 13, "y": 14}
  },
  "swing": {
    "home": {"x": 90, "y": 80},
    "range": {"x": 30, "y": 20}
  },
  "voice": {
    "lang": "ja",
    "volume": 200,
    "service": "voicetext",
    "voicetext": {
      "apiKey": "VoiceText API Key",
      "params": "speaker=hikari&speed=120&pitch=130&emotion=happiness"
    }
  },
  "chat": {
    "openai": {
      "apiKey": "OpenAI API Key",
      "model": "gpt-3.5-turbo-0613",
      "stream": true,
      "roles": [
        "Answer in Japanese.",
        "あなたはスーパーカワイイロボット「ｽﾀｯｸﾁｬﾝ」となり人々の心を癒やすことが使命です。"
      ],
      "maxHistory": 10
    },
    "random": {
      "interval": {"min": 60, "max": 120},
      "questions": [
        "何かためになることを教えて",
        "面白い話をして",
        "ジョークを言って",
        "詩を書いて"
      ]
    },
    "clock": {
      "hours": [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]
    }
  }
}

Network settings (reboot required)

• network.wifi.ssid [string] : Wi-Fi SSID
• network.wifi.pass [string] : Wi-Fi passphrase
• network.hostname [string] : mDNS hostname
• time.zone [string] : Time zone (Default: "JST-9")
• time.ntpServer [string] : NTP Server (Default: "ntp.nict.jp")

Servo/Swing settings (reboot required)

• servo.pin.x, servo.pin.y [int] : Pin number for servo (Required to swing head)
  • M5Stack CoreS3 - Port A : {"x": 1, "y": 2}
  • M5Stack CoreS3 - Port B : {"x": 8, "y": 9}
  • M5Stack CoreS3 - Port C : {"x": 18, "y": 17}
  • M5Stack Core2 - Port A : {"x": 33, "y": 32}
  • M5Stack Core2 - Port C : {"x": 13, "y": 14}
  • M5Stack Core/Fire : {"x": 21, "y": 22}
• swing.home.x, swing.home.y [int] : Home position in degrees (Default: {"x": 90, "y": 80})
• swing.range.x, swing.range.y [int] : Swing range in degrees (Default: {"x": 30, "y": 20})

Voice settings

• voice.lang [string] : Speech language for Google Translate TTS (Default: "ja")
• voice.volume [int] : Speech volume (Default: 200)
• voice.service [string] : Speech service (Default: "google-translate-tts")
  • "google-translate-tts" : Google Translate Text-to-Speech API
  • "voicetext" : VoiceText Web API
  • "tts-quest-voicevox" : TTS QUEST V3 VOICEVOX API
• voice.voicetext.apiKey [string] : VoiceText: API Key (Required to speech by VoiceText)
• voice.voicetext.params [string] : VoiceText: extra parameters (Default: "speaker=hikari&speed=120&pitch=130&emotion=happiness")
• voice.tts-quest-voicevox.apiKey [string] : TTS QUEST V3 VOICEVOX: API Key (Optional)
• voice.tts-quest-voicevox.params [string] : TTS QUEST V3 VOICEVOX: extra parameters (Default: "")

Chat settings

• chat.openai.apiKey [string] : OpenAI API Key (Required for chat)
• chat.openai.model [string] : ChatGPT model (Default: gpt-3.5-turbo-0613)
• chat.openai.stream [boolean] : Use stream or not (Default: true)
• chat.openai.roles [string[]] : Roles for ChatGPT
• chat.openai.maxHistory [int] : Send talk history (Default: 10)
• chat.random.interval.min-random.interval.max [int] : Random speech interval (Default: 60-120)
• chat.random.questions [string[]] : Questions to ChatGPT for random speech
• chat.clock.hours [int[]] : Speech hours list

API

Speak API

• Path: /speech
• Parameters
  • say : Text to speak

Example

curl -X POST "http://(Stack-chan's IP address)/speech" \
    -d "say=Hello"

Chat API

• Path: /chat
• Parameters
  • text : question

curl -X POST "http://(Stack-chan's IP address)/chat" \
    -d "text=Say something"