For AI developers
Move beyond screenshots, marketplace stars, and self-written case studies. Ask the client company to verify the project, outcome, and what may be shared.
Private beta for AI solution work
A reputation layer for AI developers companies can actually trust.
JISSEKI turns completed AI projects into company-approved proof. Strong builders earn more from verified outcomes; companies find credible AI engineers without cold-call reference chaos.
Verified AI work
01For AI developers
02For companies
03Why the beta security posture is acceptable
For companies
Evaluate AI engineers through verified work history and consented reference paths before you spend budget on a risky engagement.
Why this can scale
Like search or social ranking rewards strong content, JISSEKI is designed to reward verified AI work. Unverified claims stay weak; company-approved outcomes become discoverable proof.
Business-use validation
✓AI developers get a repeatable way to raise close rates and pricing with verified outcomes.
✓Companies reduce vendor-selection risk by seeing proof approved by a real company-domain reviewer.
✓Reviewers keep control: they approve fields and opt into future reference paths instead of exposing private contact details.
✓The loop can start from Upwork, サンカク, referrals, direct contracts, or future marketplaces.
Why the beta security posture is acceptable
No beta system should be called risk-free. For a controlled cohort, JISSEKI reduces the main abuse and privacy risks with these product and technical controls:
Company-domain verification
Project drafts require the reviewer email domain to match the company domain, reducing fake-reviewer risk.
Scoped reviewer links and OTP
Reviewers use invitation links plus OTP before submitting verification, and reviewer sessions are scoped to one request.
Hashed opaque tokens
Invitation, session, and receipt tokens are random and stored as hashes with server-side peppers, not as reusable plaintext secrets.
Consent-gated public fields
Only company-approved fields can be published. Reviewer contact details are not exposed to prospects.
Private beta access gate
Professional accounts can be restricted with BETA_ALLOWED_EMAILS, and deployment health fails when the allowlist is missing.
Database access boundaries
Supabase Row Level Security is enabled and forced on core tables; sensitive operations go through constrained server/RPC paths.
Not yet open-public-launch ready
Before broad launch, JISSEKI still needs counsel-reviewed terms, production monitoring, abuse operations, support SLAs, and security review. For known design partners, the current controls are appropriate for a small beta.
Simple beta offer
Start with one verified AI project and one friendly company reviewer. Keep the cohort small until the reference loop is proven.
Private beta
AI Developer beta
Build a verified profile from completed AI solution work.
- •Verified project record
- •Public proof page
- •Reference request inbox
Design partner
Company buyer beta
Use verified proof to find credible AI engineers.
- •Safer shortlist
- •Consent-based reference path
- •No reviewer contact exposure
How buyers start safely
This is a beta purchase path, not an open checkout yet. Start with one invited work email, one real AI project, and one company-domain reviewer before expanding spend.
- Step 1
Choose the AI developer or company-buyer beta path.
- Step 2
Sign in with an invited work email so access stays controlled.
- Step 3
Verify one completed AI solution with a company-domain reviewer.
- Step 4
Publish approved proof and route reference requests without exposing reviewer contact details.
The goal is a cleaner AI services market.
Great AI builders should become easier to find. Weak claims should become harder to sell. JISSEKI is the simple trust layer between those two outcomes.