Complete Onboarding
Create your first companion and finish the first-run setup.
Follow the bold button names and numbered steps. Technical names are included only when you need to recognize a model or file inside the app.
Onboarding introduces Local Waifu, learns your preferences, creates the first character, configures voice and models, and opens the first meeting.
Typical stages include:
- Welcome: language, AI disclosure, edition/license option.
- About You: how the companion should address you.
- Understanding You: preferences and interaction style.
- How You Connect: communication preferences.
- Your Love Language: desired expressions of care.
- Choosing: the type of dynamic you want.
- Your Dynamic: relationship tone and boundaries.
- Building Her: name, appearance, personality, and portrait details.
- Her Voice: voice preference.
- Almost There: review.
- Model Setup: local model download or Cloud AI setup.
- First Meeting: character creation is finalized and chat opens.
Configure a local model during onboarding
Goal
Install a local chat model and the semantic-memory model.
Before you start
- Connect to the internet for the model download.
- Keep Local Waifu open.
- Check available disk space.
- For a large download, prevent the computer from sleeping.
Steps
- Choose the Local AI path on Model Setup.
- Review the recommended chat model for your detected memory tier.
- Select Install and wait for the download to complete.
- Install EmbeddingGemma when it is offered. It powers semantic memory retrieval and is separate from the chat model.
- Continue after the required model is installed, or use Skip for now if the screen offers it and you intend to configure Cloud AI later.
Expected result
The chat model is ready for conversation, and the embedding model is ready to create and retrieve semantic memories.
If something goes wrong
- Check internet access, VPN, proxy, firewall, and free disk space.
- Restart Local Waifu and continue in Settings → Models.
- On Windows, allow GPU runtime setup to finish, then restart if prompted.
- If your computer cannot comfortably run the recommendation, install a smaller model in Settings → Models.
Configure Cloud AI during onboarding
Goal
Use your own provider account instead of downloading a large local chat model.
Before you start
- Have an API key or supported connected account.
- Review the provider’s pricing and privacy terms.
- Understand that selected prompts and context leave your computer when that provider is used.
Steps
- Choose the Cloud AI path.
- Select the provider.
- Paste the provider key into the secure field or complete the available account connection.
- Run the connection test.
- Select an available chat model.
- Complete onboarding.
Expected result
Chat uses the selected provider while character data, the application database, and memory storage remain local.
If something goes wrong
- Re-copy the key without spaces.
- Confirm the provider account has access and billing/credits where required.
- Turn off Offline Mode when intentionally using Cloud AI.
- Remove or replace a saved bad key in Settings → Cloud AI.