1. Setup
2. Configuration
3. Running
4. Next steps

Setup

To get Lane running you'll need at minimum an Anthropic API key. Everything else depends on which features you want enabled.

Once you have your keys, create a .env file in the root of the Wizard folder:

DISCORD_API_KEY=
ANTHROPIC_API_KEY=
DEEPSEEK_API_KEY=
QDRANT_API_KEY=
QDRANT_ENDPOINT=
OPENAI_EMBEDDING_ENDPOINT=
OPENAI_EMBEDDING_API_KEY=
AZURE_KEY=
AZURE_REGION=
ELEVENLABS_KEY=

Only populate the fields relevant to your setup — unused keys can be left blank or omitted. For full details see the .env documentation.

Configuration

Lane is configured via appsettings.json in the Wizard folder. The most important things to set for a first-time setup:

• "LLM" — set to "Claude" (default) or "DeepSeek".
• "Body" — set to "Discord" or "Terminal". If omitted, defaults to "Terminal".
• "MemoryHandlers" — if you don't have Qdrant set up, don't include a RAG handler. A SlidingWindow is sufficient to get started.
• "DefaultDiscordChannel" — required for Discord; the channel ID Lane uses for unprompted messages.

A minimal working config for Discord looks like:

{
    "Settings": {
        "MemoryHandlers": [
            {
                "Handler": "SlidingWindow",
                "ID":      "MessageWindow",
                "Args": {
                    "MaxMessages": 10,
                    "ForThoughts": 0
                }
            }
        ],
        "DefaultDiscordChannel": 0,
        "ExclusiveToChannel": false,
        "TimezoneSettings": {
            "HourShift":   0,
            "MinuteShift": 0
        },
        "RespondToThought": 5,
        "Body": "Discord",
        "Logging": {
            "ConsoleLevel": "Information",
            "FileLevel":    "Debug",
            "FileLogPath":  "lane.log"
        }
    }
}

For all configuration options including the Speech and Hearing blocks for voice, see configuration.md.

Running

Run the binary to start Lane. The body is determined by the "Body" field in appsettings.json, but can be overridden with a command-line argument:

./Wizard discord
./Wizard terminal

First run: Lane starts with no memory and builds it up as conversations happen. Memory is persisted to a data.json file next to the binary after every message, so it survives restarts.

Discord: Lane automatically joins the first voice channel in your server on startup if Speech or Hearing is configured. The monologue loop starts on connection, so Lane may send unprompted messages to DefaultDiscordChannel on her own. A dashboard TUI launches in the terminal showing live logs, token usage, next-thought countdown, and the active configuration.

Terminal: a simple read-eval loop. Type a message and press enter. The monologue does not run in Terminal mode.

Next Steps

• Customization — how to edit Lane's personality, tone, and prompt files
• How it works — the routing, monologue, and memory systems explained
• Configuration reference — all appsettings.json fields documented