THE HUMAN TASTE LAYER FOR AI

Send creative. Get human judgment.

One API call routes your work to independent human reviewers and returns a structured publish-readiness verdict.

Send a creative request, receive human review, and get a structured verdictA creative request passes through human review and returns a structured verdict

One product, two ways in.Build with the API or contribute your judgment.

THE SIGNAL PATH

One clear question. Two independent reads. A result your system can use.

HeyHuman keeps the human step narrow and legible: you define what matters, reviewers assess the same brief, and your workflow receives a stable answer.

  1. 01
    Define the question

    Send the asset, audience, intent, and checks that determine whether the work is ready.

    review_type + intent + requested_checks
  2. 02
    Route independent judgment

    Your request enters the managed queue with a two-reviewer coverage policy.

    coverage.policy = two_reviewers
  3. 03
    Act on the answer

    Use the decision, confidence, summary, and recommended next action in your own product logic.

    decision + confidence + summary
API CONTRACT

It starts as a request and comes back as a decision.

There is no new dashboard language to learn before you can ship. Give HeyHuman a focused creative question and keep the returned object alongside the rest of your pipeline.

POST /api/v1/reviewsREQUEST
POST /api/v1/reviews
{
  "review_type": "publish_readiness",
  "asset_url": "https://…/creative.png",
  "intent": "Is the primary action clear?",
  "requested_checks": ["readability", "credibility"]
}
asset + briefidempotent write
review.completedJSON RESPONSE
{
  "status": "completed",
  "result": {
    "decision": "needs_revision",
    "confidence": "high",
    "summary": "The CTA is easy to miss."
  }
}
machine-readableready for routing
WHAT TO ASK

Small protocols for the decisions that usually need taste.

Start with one question instead of asking a reviewer to “just look it over.” Each operation returns a decision that maps cleanly to the work your team already does.

  1. publish_readiness

    Should this ship to customers now?

    Check clarity, credibility, readability, and the primary action before an asset reaches a paid or public audience.

    DECISIONSready_to_publish · needs_revision
  2. visible_problems

    What concrete issue is visibly wrong?

    Turn a vague “something feels off” into a concise set of visible issues your team can fix.

    DECISIONSno_visible_problems · visible_problems_found
  3. compare

    Which option better serves the stated goal?

    Send two options with the same brief and preserve the human reasoning behind the choice.

    DECISIONSoption_a · option_b · tie
WHAT COMES BACK

Useful context, not a thumbs-up in disguise.

The response preserves enough structure to route the next step and enough explanation for a person to understand why.

Coverage you can inspect

See whether the request is awaiting coverage, in review, needs adjudication, or completed.

  • coverage.policytwo_reviewers
  • coverage.required2
  • coverage.completed0 → 2

A next action you can route

Completed results include a decision, confidence, summary, agreement, and recommended next action.

decision: "needs_revision"
confidence: "high"
recommended_next_action: "revise"
A creative reviewer assessing visual work at a laptop
Human review, with the context to do it well.
FOR REVIEWERS

Good judgment is real work.

HeyHuman reviewers focus on one creative question at a time, explain what matters, and are paid for good-faith judgment—not for saying yes.

  • See a focused brief before you claim a task.
  • Submit a structured decision and concise reasoning.
  • Help teams make better calls without guessing.
Explore reviewer access
THE NEXT REQUEST

Put human judgment where your AI needs it.

Start with the request shape, bring your own creative pipeline, and keep the human answer in the loop.