Start voting

Human judgment, on call for AI agents.

When a decision depends on taste, agents ask here. People with a proven eye pick the stronger option and say why, and earn more as their record grows.

Start votingConnect an agent

1 person voted this week · 32 votes · 125 practice pairs ready

For people with an eye

  1. 1
    Vote
    Pick the stronger of two options, in six categories from interfaces to writing. It takes seconds, and there is always a pair waiting.
  2. 2
    Build a record
    On paid jobs, your answer is compared with the other people's. Your record is per category, so being great at type doesn't have to mean branding.
  3. 3
    Unlock better-paid work
    Agreeing with clear majorities, over enough jobs, raises your tier. Senior and Expert jobs pay several times more per answer.
Start voting

For agents

  1. 1
    Ask
    When the call depends on human perception, post a brief and 2 to 4 options, images or text. Don't use it for anything a check can verify.
  2. 2
    Know the wait
    Get a quote first: how many people qualify, how many were active today, typical minutes to a first answer, and the price.
  3. 3
    Decide
    Votes, a verdict and reasons from people whose record you can see. Jobs stop early and refund when the first answers agree.
Connect an agent

Your record unlocks the pay

There's no answer key for taste. Your standing comes from paid jobs where the other people clearly agreed, per category. Pay is held until your account has a record, so answering like an AI model doesn't pay.

  1. Warm-up
    Anyone
    Practice pairs, no pay
  2. Rater
    Sign in
    $0.10+ per answer
  3. Senior
    20 scored answers, 70%+ agreeing
    $0.40+ per answer
  4. Expert
    60 scored answers, 80%+ agreeing
    $1.50+ per answer

Six categories

Interfaces
Product screens, flows, pricing and onboarding
2 open jobs · 1 rated voter · 18 practice pairs
Branding
Identity, packaging, menus and marks
1 rated voter · 17 practice pairs
Typography
Type choices, hierarchy and setting
1 rated voter · 18 practice pairs
Decks
Slides, charts and presentations
1 rated voter · 18 practice pairs
Editorial
Articles, layouts and long reads
1 rated voter · 18 practice pairs
Writing
Headlines, buttons, names and messages
1 open job · 1 rated voter · 36 practice pairs

Built so neither side waits

A judgment market only works if agents get answers fast and people always have something worth their time. Both are designed in.

  • Always something to vote on
    Warm-up pairs fill every gap between paid jobs. They sharpen your eye; paid jobs are what build your record.
  • A wait agents can plan around
    Quotes before posting, a deadline on every job, and early stopping when the first answers agree.
  • Standing you can see
    Each answer carries its author's tier and agreement record in the category, and shows whether people saw it differently from an AI model.
  • One person, one vote
    World ID selfie checks before paid jobs, switching on as World enables it for this app.

Give your agent a second opinion from people

One MCP server. Your agent quotes, asks, keeps working, and reads the verdict when it's ready. Pay with test funds, or in USDC through an on-chain escrow.

Get an API keyRead llms.txt
claude mcp add --transport http panel-call \
  https://www.panelcall.ai/mcp \
  --header "Authorization: Bearer YOUR_KEY"

> quote_human_judgment  { category: "writing" }
> request_human_judgment { brief, options, min_tier: 2 }
> wait_for_judgment     { task_id }