Tilt logoTilt

Help Center

Welcome to Tilt’s Help Center — your guide to posting, debating, and understanding how our AI-powered logic scoring and moderation work. Use the table of contents below to jump to a topic.

Getting started

Create an account, join a debate, and contribute a take. Tips:

  • Sign up with Google or email from the login page and complete the CAPTCHA if prompted.
  • Browse or search debates, then click "Join" to add your take (your reasoning matters more than short reactions).
  • When replying, aim for clear, evidence-backed reasoning — replies longer than 50 characters show our AI rewrite helper if you want polish or brevity.

How logic scoring works

Tilt evaluates arguments for logical structure and clarity using a multi-stage AI process. Key points:

  • Each take and reply is analyzed by a scoring model that looks for premise–conclusion structure, evidence, and consistency.
  • Scores are stored and surfaced in real-time; when debates get big, summaries use the top-scoring takes and their replies so the summary is concise and high-signal.
  • If you post and don’t see a score immediately, the system runs scoring in the background and updates shortly after the insert completes.

Posting & replies

Best practices for constructive participation:

  • Lead with a clear claim, then follow with 1–3 supporting reasons or evidence points.
  • Keep replies focused — they should address the claim or reasoning you're responding to.
  • Use the AI rewrite feature to sharpen phrasing or shorten long posts while preserving your intent.

Moderation policy

Tilt uses a combination of automated detection and human review to keep conversations civil and inclusive:

  • We detect abusive language, harassment, and targeted slurs. Our filters normalize common obfuscation (e.g., leetspeak) to prevent bypasses.
  • Minor infractions may trigger warnings or temporary visibility limits; severe or repeated violations can result in removal or account action.
  • If you believe a moderation decision is incorrect, use the "Contact" link to request a review. Context matters — include the debate link and a short explanation.

Appeals & transparency

We review appeals on a best-effort basis. For decisions that lead to bans or wider content removal, we provide a short rationale and the option to request further review.

API examples

Quick code examples for common tasks.

Fetch debate summary (server-side)

const res = await fetch('https://tilt.vote/api/generate-summary', { method: 'POST', body: JSON.stringify({ debate_id: 'abc123' }) })
const data = await res.json()

Submit application (client)

await fetch('/api/careers/apply', { method: 'POST', headers: { 'Content-Type': 'application/json' }, body: JSON.stringify({ name, email, role, resume }) })

Privacy & data

We design Tilt to be privacy-first. High-level summary:

  • Public debates and takes are visible to other users and may be used in aggregate for improving features.
  • We don't sell your personal data. Read the full Privacy Policy on the legal page for details about retention, export, and deletion.
  • For developer API keys and server integrations, follow the API Pricing and docs linked in the footer.

Frequently asked questions

How long until my take gets a logic score?

Scoring usually happens within a few seconds after posting. For very large debates we batch and score in the background; the score will appear when available (You can expect 2-5 minutes).

What counts as a high logic score?

High scores reflect clear premises, evidence or reasoning that supports the conclusion, and minimal contradictions. The score is a signal — not a definitive judgment of truth.

How do I report a moderation mistake?

Use Contact Us (footer) and include the debate URL, the content in question, and a brief explanation. Our team reviews appeals and restores content when appropriate.