Building with you, beyond what you imagined.
AI for the front lines,
ready to use with confidence.
Reviewing contracts, answering inquiries, finding things in internal documents.
Cor.Inc., based in Fukuoka, checks where AI can help, even in work that handles information you can't send outside, and works with you from design through operations.
When it's right, we'll even propose not building yet.
- ISMS Stage 2 audit completed (Oct 2026)
- Developed and operates Engineer Cafe's multilingual AI reception
- Born in Fukuoka, serving all of Japan
Your challenges
Let's sort out stuck work, starting from why it stopped.
We design how to choose between AI that keeps information in-house and AI tools your company has approved, based on how important the information is.
We cover everything from deciding what to build to the user experience, system integration, billing, automated testing and release, and monitoring.
From organizing the data you use today to migrating it to the cloud and designing how it runs, we move forward with you.
Using voice support, AI trained on your internal documents, and AI that runs on its own, we design guidance for facilities and local governments.
We build systems that process industry-specific data with AI, such as architectural drawings and floor-plan generation.
From shaping how you use AI to internal training and how you run development, we work alongside your team.
Proof
Our track record and safeguards, in a form you can check.
Each case page walks through the problem, design, build, and operation in order.
Engineer Cafe's multilingual AI reception
For Engineer Cafe in Fukuoka City, we developed Engineer Cafe Navigator, a multilingual voice AI reception system, and published it as open source (in live operation).
Moving a core system run for about 25 years to the cloud
Starting with research that separates facts from assumptions, we are moving an educational institution's core database to the cloud in stages.
ISMS Stage 2 audit completed
We completed the Stage 2 audit for ISO/IEC 27001 (ISMS) in October 2026 and expect certification registration around December 2026.
Fukuoka's 100 Representative Companies 2026-2027
We were selected for Made In Local's "Fukuoka's 100 Representative Companies 2026-2027" and are featured on its special website.
Five pillars that make AI ready to use with confidence.
Our own product that extracts verifiable facts from an engineer's development record. It assigns no scores and gives the same result no matter who runs it. Sales teams at recruiting agencies and SES companies can use it to explain why they recommend a candidate, backed by evidence. Hiring teams at companies that build their own software can use the records a candidate chooses to share to see what their work has actually involved.
We build AI environments that can handle information that is hard to hand to external AI, such as contracts and customer data — including the design of internal rules, permissions, and logs. We choose between cloud AI and local LLMs according to how sensitive the information is.
Confidential-data AI assessmentHardware AI | Using AI to make up for limited vision.
Together with a development partner, we are developing smart glasses for people with visual impairments. We are preparing a field trial together with users.
We work with you end to end — from framing the problem to design, implementation, and operations. Using generative AI, RAG, AI agents, image and audio processing, and more, we help you build AI systems that fit into day-to-day work, including AI usage policies and internal rules.
Through teaching at a vocational school (planned from April 2027), our work as an education partner for an international programming certification, and preparations for a generative AI exhibition in Fukuoka (being planned with Engineer Cafe), we help grow the number of people who drive AI in Kyushu.
Security
We address the worries of handling confidential data through design.
Input data, access rights, operation logs, human approval, deletion rules, and who is responsible for operations. We settle each worry about handling confidential data in the design, one by one.
- We review contracts, AI, personal information, and ISMS operations together with our retained legal counsel
- We choose between approved AI, company accounts, and isolated environments according to how sensitive the information is
- Company-issued Macs, encryption, MDM (device management), and lightweight controls centered on detection and response
By industry
Every industry gets stuck for different reasons.
How data is handled, how work flows, and how decisions are made all differ from one workplace to another. Before we start building, we sort out how to proceed with you.
- Healthcare
Organizing clinical data, patient response, information security, AI strategy
- Professional services
Contract review, applications & registration, deadline management, subsidy applications
- Construction
Labor management, design changes, safety checks, passing on know-how
- Manufacturing
Visual inspection, passing on know-how, equipment maintenance, inventory management
- Education
Test workload, parent communication, surveys & reports, ICT roster management
Our product
Résumés can be padded. Evidence can't.
Grift is our own product that extracts verifiable facts from an engineer's development record (git history). Sales teams at recruiting agencies and SES (IT staffing) companies can use it to explain why they recommend a candidate, backed by evidence. Hiring teams at companies that build their own software can use the records a candidate chooses to share to see what their work has actually involved.
No scores
It reports measured values. Judgment stays with people.
Same result every time
No AI (LLM) is used for measurement, so anyone who runs it gets the same numbers.
Sends nothing anywhere
It runs only on your own computer. Whether to share the results is up to the engineer.
Grift CLI is published free of charge as open source under the MIT License. We welcome inquiries about trials (PoCs) from recruiting agencies, SES companies and companies that build their own software.
# Install (Python 3.11+)
$ pipx install grift-cli
# Measure and record
$ grift report
# Recompute under the same conditions to verify
$ grift verify
VERIFIED: all fields match recomputation under recorded provenanceFAQ
Our business has five pillars: Grift, local LLM & secure AI, hardware AI, AI contract development & adoption support, and education & community. For AI adoption and development, you can consult us end to end — from framing the problem through design, implementation, and operations. Some of our hardware AI (smart glasses) and education & community work is still in preparation or planned.
The first decision is not which AI model to use. It is clarifying which data can be processed, which requires anonymization, and which must be kept in-house or in a local LLM. At Cor.Inc., we design input boundaries, permissions, logs, human approvals, and operational responsibilities first.
Cloud AI is ideal for public data or general writing. However, for sensitive documents like contracts, customer logs, and internal knowledge, we consider secure in-house environments or Local LLMs. Rather than dismissing Cloud AI, we utilize both based on data confidentiality and requirements.
We conduct workflow interviews, inventory tasks for AI adaptation, classify data, separate cloud vs. local processing, analyze security risks, and draft a PoC strategy. The estimated cost is ¥100k - ¥300k.
It varies with the problem you want to solve and the scope, but as a general guide: • Confidential Data AI Assessment: around ¥100k - ¥300k • Local LLM / AI Infrastructure PoC: from ¥3M for 3 months • Production build and operations/maintenance: quoted individually based on requirements Our quotes make the assumptions, risks, and scope of work clear.
We build practical requirements into the design: input data management, access rights, operation logs, human approval flows, deletion rules, and clear operational responsibility. Internally, we operate an ISMS and completed the Stage 2 audit for ISO/IEC 27001 in October 2026. We expect certification registration around December 2026.
Grift is our own product that extracts verifiable facts from an engineer's development record (git history). It assigns no scores or rankings and uses no AI (LLM) for measurement, so it gives the same result no matter who runs it, and whoever receives the output can verify it. Only when the person has named their own identity and a consent record exists does the report open with short sentences in Japanese on how that person's recorded work was done and with questions worth asking in an interview. The sentences themselves are not checked by verification; what can be verified are the numbers behind them. The CLI is published free of charge as open source under the MIT License, and we are accepting inquiries about trials (PoCs) that help sales teams at recruiting agencies and SES (IT staffing) companies explain why they recommend a candidate, backed by evidence, and help hiring teams at companies that build their own software see, from the records a candidate chooses to share, what their work has actually involved.
Start by checking the information you handle.
From sorting out the problem to building and running AI, we work with you end to end as an AI implementation company in Fukuoka.
If your requirements aren't set yet
Around ¥100k - ¥300k
If you want to try AI with data that can't go outside
From ¥3M for 3 months
For people who place or hire engineers
Using technology to carry what words alone can't. Why we started Cor.Inc. is told in our founder's story.
Read the founder's story