Lorentz/ 洛伦兹

Multi-observer fact-checking on X

Same claim. Different observers.
One invariant.

In relativity, the same event looks different from each frame of reference — but the invariants are what physics calls truth. Lorentz runs 4 frontier AIs in parallel and shows you both the disagreements and the invariant they all agree on.

同一条声明，4 个 AI 观察者，1 个不变量。这就是洛伦兹。

✅True❌False⚠️Misleading⚡Contested

How it works

  1. 1.

    @lorentz_ai on any tweet

    Reply with @lorentz_ai on the claim you want verified.

  2. 2.

    4 observers measure in parallel

    Claude, GPT-5, Gemini (with Google Search), DeepSeek — each from their own training frame.

  3. 3.

    See the invariant + the dissent

    Agreement %, dissenting verdicts, sources cited, full reasoning.

Pricing

Free

$0

Try it. 5 checks/day, 2 cheap observers.

  • · 5 checks / day
  • · DeepSeek + Gemini Flash
  • · Public result page

Pro

$9.90/mo

All 4 frontier observers, every claim.

  • · 500 checks / month
  • · Claude Opus, GPT-5, Gemini Pro, DeepSeek
  • · Google Search grounding
  • · 7-day trial

Power

$29/mo

For builders + research teams.

  • · 2,000 checks / month
  • · API access
  • · Watchlist tweets
  • · Priority queue