MISUMI AI Automation

Workflow control room for product SEO operations

7 agents 300K items target 95+ QA target

Secure Login

Sign in

MI

MISUMI AI

Automation Console

Start Projects Manual Run Blueprints History System

SaaS MVP

Start with Copilot

File-first setup Preview before write Test run first

Workspace setup

Choose the default generator before starting projects

Required
Advanced settings

Default is Codex Local. Add an endpoint now, or save and configure it later.

Start here

Let Copilot set up the work

Preview first

Safe batch runner

Use small scheduled batches so QA can pause risky full runs before the AI produces too much content.

File-first setup

CSV/XLSX source catalog file. Copilot will prepare the project plan from the actual columns.

No file reviewed yet.

Start by uploading a file or describing the workflow. Nothing is written until you approve the plan.

Plan

Review before execution

No actions

Workflow to review

  1. Upload / describeOperator gives a series file or intent.
  2. Map source fieldsA-D stay catalog evidence, not SEO output.
  3. Preview planBlueprint, project, run plan, fields, criteria, and first test run.
  4. Approve safe testGenerate a small sample before full-file processing.
  5. QA gateRows below threshold or thin evidence go to review.
  6. Scheduled full batches100-500 rows at a time, auto-pause when review rate is high.
  7. Review / exportInspect rows, then download XLSX/CSV/HTML output.
พิมพ์สิ่งที่อยากให้ AI ช่วยจัดการ แล้วกด Preview Plan

Project Management

Projects

Auto-refresh
Not sure what to do next? Start with Copilot to review a file, create the project, attach a blueprint, and run a small QA sample.
Start with Copilot

Project

Manual execution

Run plans

Use this only when you already know which blueprint and file to run. For guided setup, go back to Copilot. Pick a saved blueprint or create a new blueprint →

Quality Criteria

applies to runs in this project

Project Fields

custom metadata

Fields describe extra columns or review data this project should track. Copilot can add these from a command.

Advanced path

Manual Run is for operators who already know the file, blueprint, and run mode.

For first-time setup, let Copilot review the upload and create the project/run plan before processing.
Use Copilot instead
  1. 1Upload
  2. 2Configure
  3. 3Run
  4. 4Review & Export

Step 1

Upload series file

xlsx / csv

Columns A–D are source catalog fields: Series Code, URL, Official Series Name, and Catalog Series Description. Manual Run generates SEO output in Columns E–I.

No file uploaded yet.

Column Mapping

Series Code Official Series Name Catalog Description

Draft Blueprint

Step 2

Choose processing blueprint

The blueprint defines stages, QA rules, and the quality target for this run.

Need a custom workflow? Create one in Blueprints, then come back — your file stays loaded.

Step 3

Processing

Starting run…

Step 4

Review & Export

Score distribution (after)

Download batch HTML
Done with this file?

Blueprint Builder

Draft

Generate a blueprint from a command, or open one from the library below to edit it visually. Use it by adding it to a project.

No blueprint loaded yet.

Edit node

Agents

Saved Assets

Blueprint Library

Agent Command

Generate a blueprint

Operations

Queue & QA Monitor

Execution

Workflow Runs

Auto-refresh

System Trace

Audit Log

Use this to compare operator/AI intent against backend execution. Run logs show the current execution mode explicitly.

Row Review

Run rows

# Series Status Risk Score Decision Review note Meta title Actions

Settings

My AI Generator

Runs you start use the model and provider API key configured here — your own key first, then the workspace key, then the server .env.

API keys (per provider)

Admin

Workspace Defaults

Admin only

Defaults for users who haven't set their own model or key.

Workspace API keys (per provider)

Research / Search layer

Use offline input-only for dry runs, or API search with Serper to collect Google/SERP context and MISUMI country-page evidence before generation.

Evaluator layer

Scores Columns H/I after generation. Use the local heuristic now, or route scoring to your own API without tying the workflow to Claude or Serper.

Admin

User Login Generator

Admin only
Email Role Created

Runtime

Agent Map