Configuration¶

Skippy stores its configuration in config.json in the application directory.

Location: C:\Users\ejb71\SkippyBuddy\config.json

Complete Configuration Reference¶

{
    "session_id": "skippy-desktop",
    "sync_whatsapp": true,
    "whatsapp_number": "+1234567890",
    "show_thinking": true,
    "thinking_level": "low",
    "voice_input_enabled": true,
    "voice_output_enabled": true,
    "stt_provider": "local-whisper",
    "whisper_model": "base",
    "tts_provider": "edge",
    "edge_voice": "en-GB-RyanNeural",
    "edge_rate": "+0%",
    "tts_rate": 175,
    "tts_voice": null,
    "wake_word_enabled": false,
    "wake_words": ["hey skippy", "skippy"],
    "wake_word_sensitivity": 0.7,
    "wake_word_cooldown": 2.0,
    "mini_mode_position": [300, 274],
    "mini_mode_opacity": 0.9,
    "avatar_enabled": true,
    "avatar_size": "normal",
    "use_pipeline": true,
    "smart_routing": true,
    "default_model": "claude-sonnet-4-20250514",
    "escalation_model": "claude-opus-4-0519",
    "enable_local_fallback": true,
    "local_model": "qwen2.5:0.5b",
    "ollama_url": "http://localhost:11434",
    "pipeline_conversation_model": "mistral:7b-instruct",
    "pipeline_fast_model": "phi3:mini",
    "pipeline_enable_streaming_tts": true,
    "pipeline_enable_templates": true,
    "pipeline_enable_few_shot": true,
    "pipeline_enable_background_claude": true,
    "pipeline_temperature": 0.8,
    "pipeline_max_tokens": 600
}

General Settings¶

session_id¶

"session_id": "skippy-desktop"

The OpenClaw session identifier. This determines:

Conversation history storage
Shared context with other channels (like WhatsApp)
Model memory

Common Values:

Value	Use Case
`"main"`	Share with all OpenClaw sessions
`"skippy-desktop"`	Desktop-only session
`"skippy-work"`	Work-related conversations

sync_whatsapp¶

"sync_whatsapp": true

Enable two-way synchronization with WhatsApp.

Value	Behavior
`true`	Responses sent to WhatsApp
`false`	Desktop-only mode

Toggle: Tray menu → 📱 Sync to WhatsApp

whatsapp_number¶

"whatsapp_number": "+1234567890"

Your WhatsApp phone number for sync. Include country code.

Format: +<country><number> (no spaces or dashes)

show_thinking¶

"show_thinking": true

Display AI reasoning/thinking blocks in chat.

Value	Behavior
`true`	Shows 🧠 Thinking blocks
`false`	Only final responses shown

Toggle: Tray menu → 🧠 Show Thinking

thinking_level¶

"thinking_level": "low"

How much reasoning the AI should use.

Level	Description	Token Usage
`"off"`	No explicit thinking	Lowest
`"minimal"`	Brief reasoning	Low
`"low"`	Moderate reasoning	Medium
`"medium"`	Detailed reasoning	Higher
`"high"`	Extensive reasoning	Highest

Cost Impact

Higher thinking levels use more tokens and cost more.

Voice Input Settings¶

voice_input_enabled¶

"voice_input_enabled": true

Enable the push-to-talk microphone button.

stt_provider¶

"stt_provider": "local-whisper"

Speech-to-text provider selection.

Value	Description	Requirements
`"local-whisper"`	Offline Whisper STT	`faster-whisper` package
`"google"`	Google Speech Recognition	Internet connection

whisper_model¶

"whisper_model": "base"

Which Whisper model to use for local STT.

Model	Size	Speed	Accuracy
`"tiny"`	~75 MB	Fastest	Good
`"base"`	~150 MB	Fast	Better
`"small"`	~500 MB	Medium	Good
`"medium"`	~1.5 GB	Slower	Very Good
`"large-v2"`	~3 GB	Slowest	Best

Change: Tray menu → Voice → 🧠 Whisper Model

Voice Output Settings¶

voice_output_enabled¶

"voice_output_enabled": false

Speak AI responses aloud.

Toggle: Click 🔇/🔊 button or Tray menu → Voice → 🔊 Speak Responses

tts_provider¶

"tts_provider": "edge"

Text-to-speech engine selection.

Value	Description	Requirements
`"edge"`	Microsoft Edge neural voices	`edge-tts` package, internet
`"pyttsx3"`	Windows SAPI voices	`pyttsx3` package

edge_voice¶

"edge_voice": "en-GB-RyanNeural"

Which Edge TTS voice to use.

Popular Options:

Voice ID	Description
`en-US-GuyNeural`	Guy (US Male)
`en-US-JennyNeural`	Jenny (US Female)
`en-US-AriaNeural`	Aria (US Female, Conversational)
`en-GB-RyanNeural`	Ryan (UK Male)
`en-GB-SoniaNeural`	Sonia (UK Female)
`en-AU-WilliamNeural`	William (AU Male)

Change: Tray menu → Voice → 🗣️ Select Voice

edge_rate¶

"edge_rate": "+0%"

Adjust Edge TTS speaking rate.

Value	Effect
`"+20%"`	20% faster
`"+0%"`	Normal
`"-10%"`	10% slower

tts_rate¶

"tts_rate": 175

Words per minute for pyttsx3 (fallback TTS).

Typical Range: 125-200

tts_voice¶

"tts_voice": null

Specific pyttsx3 voice name. null uses system default.

Pipeline Settings (Local-First AI)¶

use_pipeline¶

"use_pipeline": true

Enable the local-first pipeline using Ollama models for instant responses.

pipeline_conversation_model¶

"pipeline_conversation_model": "mistral:7b-instruct"

Main Ollama model for generating responses.

Model	Description
`mistral:7b-instruct`	Good balance of quality and speed
`llama3.1:8b`	Meta's latest model
`qwen2.5:7b`	Good multilingual support

pipeline_fast_model¶

"pipeline_fast_model": "phi3:mini"

Quick model for instant acknowledgments ("DING! Let me think...").

pipeline_enable_streaming_tts¶

"pipeline_enable_streaming_tts": true

Stream audio while text is still generating.

pipeline_enable_background_claude¶

"pipeline_enable_background_claude": true

Use Claude for complex tasks (code analysis, etc.) in the background.

pipeline_temperature¶

"pipeline_temperature": 0.8

Controls response creativity (0.0 = focused, 1.0 = creative).

pipeline_max_tokens¶

"pipeline_max_tokens": 600

Maximum response length in tokens.

enable_local_fallback¶

"enable_local_fallback": true

Enable Ollama as fallback when Claude is unavailable.

local_model¶

"local_model": "qwen2.5:0.5b"

Fallback model when primary models fail.

ollama_url¶

"ollama_url": "http://localhost:11434"

Ollama API endpoint.

Smart Routing Settings¶

smart_routing¶

"smart_routing": true

Enable intelligent routing between local models and Claude.

default_model¶

"default_model": "claude-sonnet-4-20250514"

Claude model for background/escalation tasks.

escalation_model¶

"escalation_model": "claude-opus-4-0519"

Claude model for very complex tasks.

Avatar Settings¶

avatar_enabled¶

"avatar_enabled": true

Show the animated Skippy avatar.

avatar_size¶

"avatar_size": "normal"

Avatar display size.

Value	Size
`"small"`	Compact
`"normal"`	Standard
`"large"`	Big

Wake Word Settings¶

wake_word_enabled¶

"wake_word_enabled": false

Enable hands-free activation with wake words.

Requires Additional Setup

Wake word detection requires additional dependencies.

wake_words¶

"wake_words": ["hey skippy", "skippy", "hey skip", "skip"]

Phrases that activate Skippy.

wake_word_sensitivity¶

"wake_word_sensitivity": 0.7

Detection sensitivity (0.0 = strict, 1.0 = loose).

wake_word_cooldown¶

"wake_word_cooldown": 2.0

Seconds to wait between wake word activations.

Mini Mode Settings¶

mini_mode_position¶

"mini_mode_position": [300, 274]

Screen position [x, y] for mini mode window.

mini_mode_opacity¶

"mini_mode_opacity": 0.9

Mini mode window transparency (0.0 = invisible, 1.0 = solid).

Example Configurations¶

Minimal (Chat Only)¶

{
    "session_id": "main",
    "sync_whatsapp": false,
    "show_thinking": false,
    "voice_input_enabled": false,
    "voice_output_enabled": false
}

Full Voice Experience¶

{
    "session_id": "skippy-voice",
    "sync_whatsapp": true,
    "whatsapp_number": "+1234567890",
    "show_thinking": true,
    "thinking_level": "low",
    "voice_input_enabled": true,
    "voice_output_enabled": true,
    "stt_provider": "local-whisper",
    "whisper_model": "small",
    "tts_provider": "edge",
    "edge_voice": "en-US-AriaNeural",
    "edge_rate": "+10%"
}

Fast & Cheap¶

{
    "session_id": "skippy-fast",
    "sync_whatsapp": false,
    "show_thinking": false,
    "thinking_level": "off",
    "voice_input_enabled": true,
    "voice_output_enabled": false,
    "stt_provider": "local-whisper",
    "whisper_model": "tiny"
}

Maximum Intelligence¶

{
    "session_id": "skippy-smart",
    "sync_whatsapp": true,
    "whatsapp_number": "+1234567890",
    "show_thinking": true,
    "thinking_level": "high",
    "voice_input_enabled": true,
    "voice_output_enabled": true,
    "stt_provider": "local-whisper",
    "whisper_model": "large-v2",
    "tts_provider": "edge",
    "edge_voice": "en-GB-RyanNeural"
}

Editing Configuration¶

Direct Edit¶

Close Skippy
Edit config.json with any text editor
Save and restart Skippy

Some settings can be changed through the tray menu:

📱 Sync to WhatsApp
🧠 Show Thinking
Voice → 🔊 Speak Responses
Voice → Select Voice
Voice → Whisper Model

Changes are saved automatically.

Reset to Default¶

Delete config.json and restart Skippy. A new default config will be created.

Configuration Validation¶

If config.json is malformed:

Skippy logs error to logs/error_<timestamp>.log
Falls back to default settings
Overwrites invalid config with defaults

Backup Your Config

Before making changes:

Copy-Item config.json config.json.backup