Think through a problem with local AI
Begin everyday questions or document-based research with a local model, then use web search or cloud AI only when the task needs it.
Back to product
Product details
Liora detailed guide
This guide explains how to introduce Liora and move from conversation to development or creative work. You do not have to configure every capability at once. Add the models, agents, and supporting environments required for the roles you intend to use.
From first setup to a reviewed result
- Step 1
Choose the role
Decide whether the current job is conversation, research, software development, or image, video, or music creation. The role narrows down which AI and supporting tools need to be configured.
- Step 2
Select the AI and connection
Choose a local model, Codex, Claude Code, or a supported generation environment for the task. Unconfigured capabilities remain separate, so you can begin with the parts that are ready.
- Step 3
Carry the conversation into practical work
Move an idea or research result into Work Mode files, plans, tasks, Git, and the terminal, or into the production brief in Creative Mode. The next place to inspect remains within the same application.
- Step 4
Review and preserve the result
Inspect changes, tests, generated assets, and history before continuing. Keep the conversation, settings, and working record organised so the next task can build on them.
Practical use cases
Move from a development plan to verification
Open a project and work with a coding agent through planning, implementation, diff review, and tests in the same workspace.
Organise a creative brief and generation history
Prepare the purpose, mood, style, and reference material for image, video, or music work, then reuse earlier settings and results in the next variation.
What Liora brings together and what you configure separately
The Liora workspace
Liora provides the entry points and review surfaces for the work.
- Conversation and project organisation
- Work Mode and Creative Mode workspaces
- Settings, history, diagnostics, and backup
- Fallback when an optional capability is not configured
Setup for the roles you use
Each external engine or service has its own installation or account requirements.
- Ollama or another local provider plus a model for local chat
- Codex CLI or Claude Code CLI plus sign-in for coding-agent work
- Supported software, models, or hardware for speech and generation
- A network connection and account for web search or cloud AI
Capabilities and requirements
What changes with this product
Use the right AI without rebuilding your workspace
Use local AI for everyday questions, then move to Codex or Claude Code for code analysis and implementation. The entry point stays familiar even when the work becomes more technical.
Keep conversation and practical work together
Conversation history, long-term memory, and document search sit alongside Work Mode, where files, plans, tasks, Git, and a terminal remain part of the same working context.
Automate work without removing review
File edits and external tools follow the confirmation rules you set. You can decide what the AI may do and where a person should inspect the result before continuing.
The full capability set
Conversation, search, and knowledge
Start with chat, then bring in earlier context and information stored on the PC.
- Switch between Ollama, LM Studio, and llama.cpp models.
- Voice input, speech recognition, and speech output through configured engines.
- Long-term memory, semantic search, and document search.
- Web search when current external information is needed.
Work Mode
Bring the information and controls needed for software work into one workspace.
- Browse and edit project files.
- Review plans, tasks, and proposed changes.
- Run and resume Codex or Claude Code sessions.
- Use Git and the terminal to inspect and verify work.
Creative Mode
Connect supported local generation environments and keep creation entry points and assets together.
- Entry points for image, video, and music workflows.
- Connections to supported environments such as ComfyUI.
- Installed-model checks and an asset library.
- Safe fallback when an optional generation tool is unavailable.
Extensions and maintenance
Add capabilities deliberately and retain ways to diagnose or recover the application.
- MCP servers, skills, and permission-reviewed plugins.
- PC actions and automation tasks.
- Diagnostic logs, database checks, and backups.
- Safe Mode that starts without optional external services.
Requirements and support
- Operating system
- Windows 10 or 11.
- Core AI
- Ollama is the core local provider, with connections for LM Studio and llama.cpp. Models are installed separately.
- Coding agents
- Codex CLI and Claude Code CLI are supported when each CLI is installed and signed in.
- Optional capabilities
- Speech, generation, MCP, and plugins depend on the supporting software and models you configure.
- Local data
- Conversations, memories, settings, and logs are stored in Liora's local Windows application data.
Frequently asked questions
Can I use Liora without an internet connection?
Core chat and local features can run offline once Ollama and the required models are installed. Web search, cloud AI, and external services require a network connection.
Do I need every AI and supporting application?
No. Start with Ollama and add only the capabilities you need. Missing speech, generation, or coding-agent tools are disabled without preventing the rest of Liora from working.
Can Liora run Codex and Claude Code?
Yes, after the corresponding CLI is installed and signed in. Available models and usage limits are governed by each CLI and account.
Will AI change project files automatically?
Work Mode follows its confirmation settings. Review important diffs, tests, and Git changes before accepting the result.
Can I check whether my PC is ready?
Liora includes diagnostics for model providers, speech, generation tools, and coding agents, so each optional capability can be checked separately.
Who is Liora recommended for?
Liora is recommended for people comfortable with basic Windows tasks, installing software, and downloading and configuring AI models. A core usage guide and AI setup guide are included, but individual installation work and support for PC-specific setup are not included.
Where can I buy Liora?
After reviewing the pre-purchase checklist on the official SAINO site, choose SAINO checkout powered by Stripe or Gumroad. Use the enquiry form if you need clarification before purchase.
Online delivery
Liora is available for $199 USD
Liora is provided for $199 USD as one product that includes the application, core usage guide, and AI setup guide. After reviewing the pre-purchase checklist, choose SAINO checkout powered by Stripe or Gumroad.
Price and included materials
01
Liora
$199 USD
Japan: JPY 29,800 including tax
For people comfortable configuring PCs and local AI
Receive Liora together with guidance for preparing local AI providers and coding agents. The guide does not guarantee successful setup on every PC.
- Liora installer
- Core usage guide
- Pre-installation checklist
- Codex setup instructions
- Claude Code setup instructions
- Troubleshooting guide
- Licence and release notes
Confirm these three points before checkout
From purchase to installation
- 01
Review requirements and terms
Confirm the Windows and hardware requirements, external AI setup, unsigned-installer notice, and support boundary.
- 02
Choose where to pay
After accepting the pre-purchase confirmations, choose SAINO checkout powered by Stripe or Gumroad. SAINO does not store your card details.
- 03
Receive the product
SAINO checkout uses a time-limited private link. Gumroad provides Liora, its checksum, and the setup guide from the purchase page or receipt email.
- 04
Verify and install
Compare the downloaded file with the supplied SHA-256 checksum, then follow the guide to prepare the AI providers you want to use.
The price is $199 USD. Applicable taxes may be added based on your location. Fees for external AI services, models, and paid third-party services are not included. Cancellation and refund conditions are shown before payment. The current installer is not code signed, so Windows may display a security warning.
Let’s turn an idea into something real.
Ask about a product or a custom website, web application, desktop application, or software project. Early ideas are welcome.