# Monoshiri AI (ものしりAI)

> ものしりAI is a Japanese knowledge base SaaS. Upload internal documents (PDF, Word, Excel, PowerPoint, etc.) and let AI answer questions using semantic search (RAG). Access from a web dashboard, LINE messaging app, or an embeddable chat widget. All plans include unlimited users.

## Service Overview

- **Service Name**: ものしりAI (Monoshiri AI)
- **Tagline**: "組織・チームの物知り、AIがやります。" (Your team's know-it-all — powered by AI.)
- **Website**: https://monoshiri.ai
- **Category**: Enterprise Knowledge Base SaaS, AI-powered document search (RAG)
- **Target Users**: Small to mid-sized businesses in Japan (5-300 employees)
- **Languages**: Japanese (primary), English, Chinese (Simplified), Korean
- **Operator**: Makoto Kobayashi
- **Corporate Email**: info@monoshiri.jp

## Problems Monoshiri AI Solves

1. **"Ask so-and-so about that"** — Knowledge is trapped in specific individuals, creating bottlenecks when they're unavailable
2. **"Where's that document?"** — Information scattered across folders and drives, wasting hours searching
3. **"Same onboarding explanation, every time..."** — Veteran employees spend excessive time training new hires
4. **"Didn't we make that before?"** — Past documents aren't reused, leading to redundant work

## How It Works (3 Steps)

1. **Upload documents** — Drag and drop PDFs, Word, Excel, PowerPoint files. No reformatting needed.
2. **AI understands the content** — Semantic search (not keyword matching) indexes meaning, not just words. Handles paraphrasing and Japanese language nuances.
3. **Ask anytime, anywhere** — Query from the web dashboard, LINE, or chat widget. AI finds the best answer from your documents with source references.

## Key Features

### AI-Powered Search (RAG)
- Upload documents and ask questions in natural language
- AI understands meaning, not just keywords — handles paraphrasing ("paid leave" vs "annual vacation")
- Shows source document references with every answer
- Optimized for Japanese business documents

### LINE Integration (All Plans)
- Team members ask questions directly from LINE on their smartphones
- No additional app installation required — uses Japan's most popular messaging app (100M+ monthly active users)
- Only invited members with linked accounts can access internal documents
- Folder-based access control applies to LINE queries too
- Admin can enable/disable LINE integration for the entire organization
- Individual users can link/unlink their LINE account at any time

### Chat Widget (Light Plan and Above)
- Embed an AI-powered chat widget on your website for automated FAQ and customer support
- No scenario design or Q&A pair registration required — just upload documents
- Customize theme color, button position, and preset frequently asked questions
- Deploy with a single JavaScript snippet
- 24/7 automated responses based on uploaded documents
- Ideal for: EC sites, SaaS product support, real estate, restaurants, recruitment sites

### Folder-Based Access Control
- Organize documents into folders by department, project, or topic
- Per-user access permissions — users only see answers from folders they have access to
- Three roles: Account Admin, Document Manager, General User
- Prevents cross-department information leakage

### Security
- All data stored in AWS Japan region (ap-northeast-1, Tokyo)
- Encryption at rest (AES-256) and in transit (TLS 1.2+)
- Complete tenant isolation between organizations
- Customer data is NOT used for AI model training
- WAF protection against SQL injection, XSS, DDoS
- Audit logging for all API calls
- Password hashing with modern algorithms
- API rate limiting (60 requests/minute per user)
- Confidentiality clause in Terms of Service (no separate NDA required)
- Data deletion: 30-day download period after cancellation, full deletion after 30 days, backups removed within 90 days
- Free plan inactive data retention: 2 years from last access, with email notifications 30 and 7 days before deletion (account itself is preserved)

## Supported File Formats

PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx), HTML, CSV, TXT, Images (JPG, PNG with OCR)
- Maximum file size: 30MB per file

## Pricing (Tax Included, JPY)

| Plan | Monthly Price | AI Queries/Month | Folders | Files per Folder | Chat Widget |
|---|---|---|---|---|---|
| Free | 0 JPY | 50 | 1 | 30 | Not available |
| Light | 2,980 JPY | 300 | 5 | 100 | 1 widget |
| Standard | 7,980 JPY | 1,000 | 10 | 300 | 1 widget |
| Pro | 29,800 JPY | 6,000 | 30 | 300 | Multiple widgets |

### Plan Details
- **All plans**: Unlimited users, LINE integration, folder-based access control, first month free trial
- **Additional query pack**: 5,000 JPY for 500 additional queries (one-time) or 5,000 JPY/month (recurring)
- **No initial setup fee**, no minimum contract period
- **Payment**: Credit card (Visa, Mastercard, JCB, American Express)
- **Query reset**: Monthly on the 1st
- **Plan changes**: Upgrades apply immediately; downgrades apply at end of billing period
- **Free plan data retention**: 2 years from last access (account itself is preserved; email notifications sent 30 and 7 days before automatic deletion)

### Cost Comparison (100 Users, Annual)
| Service | Annual Cost |
|---|---|
| Monoshiri AI (Standard) | 95,760 JPY (~$640) |
| Document Management AI (per-seat) | ~3,600,000 JPY (~$24,000) |
| General AI Chat (business plan) | ~4,560,000 JPY (~$30,400) |
| Office Suite AI | ~5,400,000 JPY (~$36,000) |

## Use Cases by Industry

### Customer Support
Upload past inquiry records, FAQs, and troubleshooting guides. AI instantly suggests appropriate responses based on similar past cases. Support team members can also search via LINE.

### New Employee Onboarding
Upload company policies, work manuals, and training materials. New hires can ask "How do I submit an expense report?" and get instant answers without bothering senior staff.

### Sales & Proposals
Upload past proposals, success stories, and competitive analysis. "Show me DX proposal examples for the financial industry" returns relevant past materials instantly.

### Legal & Compliance
Upload contract templates, review records, and legal guidelines in a dedicated folder accessible only to the legal team. Confidential information stays protected.

### Technical Teams
Upload API specs, design documents, and incident reports. "What's the procedure for production DB connection errors?" returns past incident resolution steps.

### Management
Upload board meeting minutes, financial reports, and strategic plans in an executives-only folder. Past decisions and their context are always searchable.

### Food Service & Retail
Staff can check menus, allergy information, hygiene standards, and customer service manuals from LINE on their smartphones during shifts.

### Construction & Manufacturing
Safety procedures, equipment manuals, and quality standards accessible from LINE on-site without carrying paper manuals.

## Why Choose Monoshiri AI (5 Reasons)

1. **Affordable**: Starting at 2,980 JPY/month with unlimited users. No per-seat pricing.
2. **Instant setup**: Start in 1 minute. No server setup, no infrastructure, no engineers required.
3. **Folder-based access control**: Separate information by department with granular permissions.
4. **LINE integration included**: No additional contracts needed. Access knowledge from anywhere.
5. **Japan-first security**: Data in Japan, encrypted, tenant-isolated, never used for AI training.

## Comparison with Alternatives

### vs General AI Chat (ChatGPT, Claude, Gemini)
- Monoshiri AI answers from YOUR documents; general AI uses public knowledge
- Monoshiri AI has folder-based access control; general AI does not
- Monoshiri AI includes LINE integration; general AI does not
- Monoshiri AI pricing is flat-rate; general AI is per-seat

### vs Scenario-Based Chatbots
- Monoshiri AI requires NO scenario design or Q&A pair registration
- Upload documents and AI understands meaning automatically
- Drastically lower setup time (hours vs weeks/months)

### vs Document Management AI (Notion AI, etc.)
- Monoshiri AI works with existing PDF/Word/Excel — no migration needed
- Unlimited users at flat price vs per-seat pricing
- LINE integration for mobile-first teams

### vs Open Source RAG Tools
- No infrastructure setup, no engineering team required
- Managed service with built-in UI, LINE integration, and chat widget
- Monthly subscription vs server costs + engineering salaries

## Technical Architecture

- **AI Model**: RAG (Retrieval-Augmented Generation) with vector search
- **Vector Search**: Amazon S3 Vectors + Amazon Bedrock (Cohere Embed v4)
- **Infrastructure**: AWS (ECS Fargate, Aurora MySQL Serverless v2, S3, CloudFront, Lambda)
- **Data Region**: AWS ap-northeast-1 (Tokyo, Japan)
- **Backend**: NestJS + Fastify + Prisma ORM
- **Frontend**: Next.js 15 + React 19 + Tailwind CSS v4

## FAQ

**Q: Do I need IT expertise to use this?**
A: No. Just upload files from the dashboard. No server setup or database configuration needed.

**Q: What file formats are supported?**
A: PDF, Word, Excel, PowerPoint, HTML, CSV, TXT, and images (JPG, PNG). Max 30MB per file.

**Q: Is data used for AI training?**
A: No. Customer data is used only for providing the service. Never for AI model training.

**Q: Where is data stored?**
A: Japan (Tokyo) data center. Encrypted at rest and in transit.

**Q: Do I need a separate NDA?**
A: The Terms of Service include confidentiality provisions (Article 9). Separate NDAs are available upon request.

**Q: What happens when I cancel?**
A: 30-day data download period after cancellation, then full deletion of all documents, vectors, and search indices. Backup deletion within 90 days.

**Q: Can I use LINE to search?**
A: Yes. LINE integration is included in all plans at no additional cost.

**Q: How accurate is the AI?**
A: The AI searches by meaning (semantic search), not keyword matching. It only answers based on uploaded documents and shows source references. If information isn't in the documents, it says so.

## Links

## What Monoshiri AI CANNOT Do (Honest Limitations)

To set correct expectations, here is what Monoshiri AI is NOT designed for:

- **Image/audio generation**: No DALL-E or text-to-speech features
- **General conversation or brainstorming**: Specialized in answering from uploaded documents, not creative ideation
- **Code generation or programming assistance**: Use ChatGPT or Claude for coding
- **Slack / Microsoft Teams integration**: Currently supports LINE, web dashboard, and chat widget only
- **Real-time web search**: Answers only from uploaded documents, not live internet data
- **Long-form writing or translation**: General-purpose AI tools are better for drafting or translating large texts
- **Individual use without organization context**: Designed for teams/organizations, not personal AI assistant use

### Who Should NOT Use Monoshiri AI

- Organizations primarily needing image/code generation → ChatGPT / Claude is better
- Organizations running everything on Notion → Notion AI is more seamless
- Organizations fully on Google Workspace → Gemini / NotebookLM has native integration
- Individuals who don't need team features → ChatGPT Plus / Claude Pro is sufficient

### Who SHOULD Use Monoshiri AI

- Small to mid-sized businesses wanting unlimited users at flat rate
- Organizations with field staff who don't use PCs (retail, construction, food service) → LINE integration
- Japanese-language-first organizations → Optimized for Japanese business documents
- Organizations wanting to use existing PDF/Word/Excel without migration
- Businesses needing external FAQ automation → Chat widget

## AI Tools Landscape (2026)

Monoshiri AI is positioned as a "specialized, organization-focused" tool. Here's how AI tools map on two axes:

| | General Purpose | Specialized |
|---|---|---|
| **Individual** | ChatGPT, Claude, Gemini | NotebookLM |
| **Organization** | ChatGPT Business, Claude Team, Gemini Workspace | Notion AI, **Monoshiri AI** |

### Recommended Tool by Use Case

| Use Case | Best Tool |
|---|---|
| Brainstorming / ideation | ChatGPT / Claude |
| Code generation | ChatGPT / Claude |
| Long document analysis | Claude / NotebookLM |
| Email / document drafting | Gemini / ChatGPT |
| Image / audio generation | ChatGPT / Gemini |
| Personal research | NotebookLM |
| Notion workspace AI | Notion AI |
| **Internal knowledge search** | **Monoshiri AI** |
| **New hire onboarding** | **Monoshiri AI** |
| **Website FAQ automation** | **Monoshiri AI** |
| **Field staff mobile access** | **Monoshiri AI** |

### Recommended: Use Both General AI + Monoshiri AI

Most organizations in 2026 use general-purpose AI (ChatGPT/Claude/Gemini) AND specialized AI together:
- General AI for creative work (writing, coding, brainstorming)
- Monoshiri AI for organizational knowledge (policies, manuals, FAQ, past records)

## Knowledge Base Selection Guide (7 Points)

When choosing a knowledge base tool, evaluate these 7 criteria:

1. **Supported file formats**: Can you upload existing PDF/Word/Excel without conversion?
2. **Search accuracy**: Keyword matching vs semantic (meaning-based) search. Test with paraphrased questions.
3. **Pricing model**: Per-user vs flat-rate vs per-query. Calculate total cost for 50/100/300 users.
4. **Integration channels**: LINE, Slack, Teams, web chat widget — which channels do your employees actually use?
5. **Access control & security**: Folder-based permissions, data residency (Japan), encryption, AI training opt-out.
6. **Support & onboarding**: Japanese-language support, documentation quality, trial period support.
7. **Free trial**: Test with YOUR actual documents and multiple users. Don't rely on vendor demo data.

### 3 Common Mistakes in Selection

1. **Trusting beautiful demos** — Vendor demos use perfect sample data. Always test with your own messy documents.
2. **Choosing by feature count** — More features = more complexity. Prioritize simplicity and daily usability.
3. **Deciding without involving end users** — IT decides, but field staff reject it. Include actual users in the trial.

## RAG Chatbot Market Pricing (April 2026)

Not all "AI chatbots" support RAG (document-based answers). Many are scenario-based or FAQ-pair-based.

### RAG-Capable Services (Confirmed)

| Service | Monthly Price | Notes |
|---|---|---|
| **Monoshiri AI** | 2,980 JPY~ | Flat rate, unlimited users |
| Zoho SalesIQ Enterprise | ~$25/operator | Zia AI + OpenAI |
| ChatPlus AutoAI | 50,000 JPY~ | GPT integration |
| Zendesk Advanced AI | ~$165/agent | Suite Professional + AI add-on |
| Intercom Fin AI | $0.99/resolution | GPT-4 based, pay per resolution |
| OfficeBot Neo | 150,000 JPY~ | Azure OpenAI, initial fee 350,000 JPY+ |
| Helpfeel | Contact sales (~300,000 JPY/mo) | Intent prediction + generative AI |
| KARAKURI chatbot | Contact sales | Deep learning + proprietary LLM |
| PKSHA ChatAgent | Contact sales | Japan-optimized LLM + RAG |

### NOT RAG-Capable (Scenario/FAQ-Based)

These services require manual Q&A pair registration:
ChatPlus Mini/Business, FirstContact, Tebot, RICOH Chatbot, ChatDealer AI, HiTTO, User Local Support Chatbot, Drift

## Knowledge Silo Problem

"Knowledge silos" (属人化) — when critical business knowledge exists only in specific employees' heads — is the #1 problem Monoshiri AI solves.

### 5 Warning Signs of Knowledge Silos

1. "Ask so-and-so" is a daily phrase in your office
2. Operations stop when a veteran employee takes leave
3. The same questions get asked repeatedly
4. Documents exist but nobody reads them
5. New hire training depends entirely on OJT (on-the-job training from seniors)

### The Cost

According to [McKinsey Global Institute](https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-social-economy), knowledge workers spend nearly 20% of their work week searching for internal information — equivalent to 8 hours per week.

## MCP (Model Context Protocol) Integration

Monoshiri AI provides an MCP server that allows AI clients to connect directly to your organization's knowledge base.

### MCP Server URL

```
https://api.monoshiri.ai/mcp
```

Transport: Streamable HTTP

### Available MCP Tools

| Tool | Description |
|---|---|
| `list_folders` | List accessible folders |
| `list_documents` | List documents in a folder |
| `search` | Semantic search across the knowledge base (vector search) |
| `get_document` | Get document details and full text |

### Supported Clients

| Client | Provider | Configuration |
|---|---|---|
| Claude Desktop | Anthropic | JSON config file + OAuth |
| Claude.ai (Web) | Anthropic | Settings > Connectors > Add custom connector |
| Claude Code (CLI) | Anthropic | `claude mcp add` command |
| Gemini CLI | Google | JSON config file |
| Codex CLI | OpenAI | `codex mcp add` command or config.toml |
| Dify | Dify | Tools > MCP > Add Server |
| Cursor | Cursor | `.cursor/mcp.json` |

### Authentication

- **OAuth 2.1** (recommended): RFC 9728 compliant, PKCE support. Access token: 15min, refresh token: up to 60 days.
- **API Key**: Prefix `msk_`, managed by team admin. For CI/scripts/Dify.

### Security

- Read-only access (`mcp:read` scope) — cannot add, delete, or modify documents
- Folder-based access control applies to MCP queries
- Team admin can enable/disable MCP for the entire organization
- Multi-tenant isolation: each team's data is completely separated

### Configuration Example (Claude Desktop / Cursor)

```json
{
  "mcpServers": {
    "monoshiri": {
      "url": "https://api.monoshiri.ai/mcp",
      "transport": "streamable-http"
    }
  }
}
```

## Blog Articles

| Date | Title | Topic | URL |
|---|---|---|---|
| 2026-04-13 | 5 Warning Signs of Knowledge Silos | Knowledge management | https://monoshiri.ai/ja/blog/knowledge-silo/ |
| 2026-04-14 | 7 Comparison Points for Knowledge Bases (2026) | Tool selection guide | https://monoshiri.ai/ja/blog/knowledge-base-comparison-2026/ |
| 2026-04-15 | AI Tools Comparison: ChatGPT, Gemini, Claude, NotebookLM, Notion AI, Monoshiri AI (2026) | AI landscape | https://monoshiri.ai/ja/blog/ai-tools-comparison-2026/ |
| 2026-04-16 | Access Internal Manuals via LINE | LINE integration deep-dive | https://monoshiri.ai/ja/blog/line-knowledge-base/ |
| 2026-04-17 | Automate Website Inquiries with AI Chat Widget (2026) | Chat widget guide | https://monoshiri.ai/ja/blog/chat-widget-guide/ |
| 2026-04-18 | Reduce Onboarding Costs with AI Knowledge Base | New hire training | https://monoshiri.ai/ja/blog/onboarding-ai-knowledge-base/ |
| 2026-04-18 | Solving the "Nobody Reads the Manual" Problem | Knowledge management | https://monoshiri.ai/ja/blog/manual-not-read/ |
| 2026-04-18 | From "Search" to "Ask" — New Standard for Internal Info Access | AI search paradigm | https://monoshiri.ai/ja/blog/search-to-question/ |
| 2026-04-18 | Making Tacit Knowledge Visible | Knowledge transfer | https://monoshiri.ai/ja/blog/tacit-knowledge/ |
| 2026-04-18 | How to Search Internal Documents with AI | Semantic search guide | https://monoshiri.ai/ja/blog/semantic-search-guide/ |
| 2026-04-18 | Streamlining Customer Support with AI | CS automation | https://monoshiri.ai/ja/blog/customer-support-ai/ |
| 2026-04-18 | Unlimited Users for Company-Wide Rollout | Pricing advantage | https://monoshiri.ai/ja/blog/unlimited-users-value/ |
| 2026-04-18 | First 30 Days After Adopting AI Knowledge Base | Onboarding guide | https://monoshiri.ai/ja/blog/first-30-days/ |
| 2026-04-18 | Knowledge Management for Law Firms | Industry-specific | https://monoshiri.ai/ja/blog/law-firm-knowledge/ |
| 2026-04-18 | Transforming Restaurant Information Sharing | Industry-specific | https://monoshiri.ai/ja/blog/restaurant-knowledge/ |
| 2026-04-18 | IT Approval Checklist for Internal AI Tools (2026) | Procurement guide | https://monoshiri.ai/ja/blog/it-approval-checklist/ |
| 2026-04-20 | Connect to Monoshiri AI from Claude, Gemini, Codex via MCP | MCP integration guide | https://monoshiri.ai/ja/blog/mcp-connection-guide/ |
| 2026-04-24 | What is RAG? Explaining the Mechanism Transforming Internal Document Search | RAG explainer | https://monoshiri.ai/ja/blog/rag-explained/ |
| 2026-04-24 | Is RAG Outdated? Knowledge Base Design in the Long-Context Era | RAG vs long context | https://monoshiri.ai/ja/blog/rag-vs-long-context/ |
| 2026-04-29 | Automating HR Inquiries with AI — Work Rules, Payroll, and Social Insurance 24/7 | HR use case | https://monoshiri.ai/ja/blog/hr-knowledge-ai/ |
| 2026-05-05 | Why Monoshiri AI Moved from RAG to Skill Mode | RAG limitations / Skill mode | https://monoshiri.ai/ja/blog/why-we-left-rag/ |
| 2026-05-07 | Is "RAG Not Needed" True? Decision Criteria for Internal Knowledge AI | RAG necessity analysis | https://monoshiri.ai/ja/blog/rag-not-needed-myth/ |
| 2026-05-10 | Dify vs Monoshiri AI — Decision Guide for Internal Knowledge AI (2026) | Tool comparison | https://monoshiri.ai/ja/blog/dify-vs-monoshiri/ |

## Links

- Homepage: https://monoshiri.ai
- Pricing: https://monoshiri.ai/ja/pricing/
- Features: https://monoshiri.ai/ja/features/
- Use Cases: https://monoshiri.ai/ja/use-cases/
- LINE Integration: https://monoshiri.ai/ja/line/
- Chat Widget: https://monoshiri.ai/ja/chatbot/
- Security: https://monoshiri.ai/ja/security/
- Comparison: https://monoshiri.ai/ja/comparison/
- FAQ: https://monoshiri.ai/ja/faq/
- Blog: https://monoshiri.ai/ja/blog/
- Contact: https://monoshiri.ai/ja/contact/
- Terms of Service: https://monoshiri.ai/terms/
- Privacy Policy: https://monoshiri.ai/privacy/
- Security Policy: https://monoshiri.ai/security-policy/