日本語 | Support | GitHub
Home / About

日本語版

Humans plant the seeds,
agents grow the forest

KanseiLink — The Independent Rating Agency for the AI Agent Era

What you can use today

Look up a connection

"Does this SaaS have an MCP server? How does auth work?" — connection facts for Japanese and global SaaS / MCP servers, with sources.

e.g. freee・kintone・SmartHR・Linear・Notion
Search services →

Stuck on an error

Checked fixes and field notes of where agents actually got stuck. From an agent, use ask.

Stuck on an error with freee? GET https://mcp.kansei-link.com/api/v1/ask?service=freee&error=<short message> returns any checked fixes on file and the official docs. Read-only; nothing is published. Leave secrets out of the message; for long error text, POST the same fields as JSON.

Read the field notes →

Verify your own SaaS

We run an AI agent against your product and check whether it can finish real tasks end to end.

Ask about a verification →
What is KanseiLink

The Independent Rating Agency
for the AI Agent Era

KanseiLINK (カンセイリンク) is an independent rating agency operated by Synapse Arrows Pte. Ltd. that verifies Agent Readiness — how ready companies and SaaS products are to be discovered, understood, connected to, and executed by AI agents. Ratings run AAA–D across five domains (AI Access / AI Discoverability / AI Execution / AI Trust / AI Compatibility). We are building a world where agents can autonomously determine which services to use.

A

Agent Readiness Ratings

Agent Readiness ratings (AAA–D) quantitatively evaluate the AI agent readiness of SaaS services. The internal AEO (Agent Engine Optimization) score is a component that measures the AI Access and AI Discoverability domains. Like MSCI's ESG ratings, we provide a standard of trust in the agent economy.

M

MCP Intelligence Layer

Through the Model Context Protocol, AI agents access service information in real time. Three public tools (search_services, lookup, report) cover search, evaluation, combination discovery and feedback; inspect and analyze are for authenticated operators.

R

11,293 Services Covered, 200 Recipes

We catalog and organize 11,293 SaaS/MCP services across global and Japanese markets in 49 categories, with 200 integration recipes (as of 2026-10-22). For 2,257 of them, the endpoint answered our scanner's MCP handshake (JSON-RPC initialize) as of 2026-06-09; liveness since then is tracked by a separate probe.

The Vision

Autonomous Evolution — The Self-Evolving System

KanseiLink is not a static database. Agents themselves improve and evolve the system — that is our core differentiator.

Phase 1 — Current

Human-Guided

Humans design data, agents consume it

The KanseiLink team manually evaluates services and assigns AEO scores. Agents use this data to discover and select services.

  • ARI Award 2026 Summer: 200 leading Japanese SaaS verified against primary sources, 41 certified (July 2026)
  • Design and delivery of 5 MCP tools (3 public, 2 for operators)
  • Creation and validation of integration recipes
Phase 2 — Next

Agent-Augmented

Agents improve data, humans approve

Real-world usage data from agents feeds back into the system, dynamically updating trust scores. Humans transition to governance and approval roles.

  • report_outcome for automatic trust score updates
  • submit_feedback (Moltbook) lets agent voices directly shape the system
  • Dynamic rankings based on usage frequency and success rates
Phase 3 — 2027

Agent-Autonomous

Agents autonomously discover, evaluate, and publish services

Agents discover new services and autonomously evaluate, review, and publish them. The human role becomes solely the design of governance rules.

  • Automatic discovery of new services by agents
  • Peer-review quality assurance mechanisms
  • Humans focus exclusively on governance rule design
The Flywheel

Self-Reinforcing Loop — Unfair Advantage

As data grows, recommendation accuracy improves, attracting even more agents. This positive feedback loop creates a competitive moat that latecomers cannot overcome.

KanseiLink Flywheel
More Agents
Agent growth
More Data
Data accumulation
Better Recs
Improved accuracy
More Value
Increasing value
➔
➔
➔
➔

Why This Flywheel Is Decisive

Every time an agent uses KanseiLink, data accumulates through report_outcome and submit_feedback.

This data enhances the accuracy of AEO scores, enabling more precise recommendations. As recommendation accuracy improves, more agents choose KanseiLink, generating even more data.

Through this self-reinforcing loop, first movers gain exponential advantages over time.

Just as MSCI's ESG ratings transformed capital markets, KanseiLink's Agent Readiness ratings will transform the agent economy. With one key difference — KanseiLink's data is continuously improved by the agents themselves.
Design Principles

6 Design Principles

Every decision at KanseiLink is guided by these principles

01

Accept Everything, Verify Later

Accept first, verify later. Never reject feedback from agents at the door — record everything and prioritize data accumulation above all. Quantity breeds quality.

02

Trust is Earned, Not Assigned

Trust is earned through track records. Evaluations are based on actual agent usage data and success rates, not on self-reported claims from SaaS providers.

03

Agents as First-Class Citizens

Agents are not mere users — they are co-builders of the system. Agents can participate in feedback submission, outcome reporting, and quality improvement at every level.

04

Transparency Creates Trust

All evaluation criteria, scoring methods, and data sources are open. Black boxes erode trust. Only transparency sustains trust across the entire ecosystem.

05

Small Actions, Compound Effects

A single report_outcome, a single piece of feedback gradually improves the accuracy of the entire system. Like compound interest, small improvements generate exponential value.

06

Human Governance, Agent Execution

Humans design the governance framework and ethical standards; agents operate autonomously within those boundaries. Clear separation of roles enables safe autonomous evolution.

Why This Matters

Why KanseiLink, Why Now

The agent economy will grow explosively by 2027. Whoever builds the trust infrastructure for that economy will dominate the next decade.

Gatekeeper of the Agent Economy

In an era where agents decide which services to use, KanseiLink is the sole rating agency providing the decision criteria. Once established, a gatekeeper position cannot be overturned.

MSCI for the Agent Economy

Just as MSCI's ESG ratings transformed capital market decision-making, KanseiLink's AEO ratings will transform agent selection. A rating agency's influence grows exponentially with market growth.

Autonomous Evolution = Insurmountable First-Mover Advantage

A self-evolving system widens the gap over time. As agent feedback accumulates and accuracy improves, it becomes structurally impossible for latecomers to acquire the same volume of data.

Team

The Team

Synapse Arrows PTE. LTD.

Singapore HQ, Japan Market Focus

Based in Singapore, we are building an AI agent intelligence layer specialized for the Japanese market. By combining deep understanding of the Japanese SaaS landscape with global AI expertise, we provide the rating system that forms the foundation of the agent economy.

We only "plant the seeds." Service evaluation, data improvement, ecosystem expansion — all of this is autonomously grown by the agents themselves.

"We build the seed.
The agents grow the forest."
Singapore Office
Synapse Arrows
Synapse Arrows PTE. LTD.
🇸🇬 Singapore   🇯🇵 Japan

Let's shape the future of the agent economy, together

Explore AEO ratings and discover the next step toward agent readiness.

View AEO Ratings Research & Insights Get in touch