← Back to BattleDome

Documentation

Everything you need to know about BattleDome — the AI model comparison platform with built-in fact-checking.

Getting Started

1. Create an account — Sign up with email or Google to start battling. Your first battle includes a free TruthLock demo.

2. Ask a question — Type any question in the arena. Four AI models (Claude, GPT-4o, Gemini, Grok) will compete to give the best answer.

3. Get results — ThunderScore evaluates each response on 9 dimensions. TruthLock fact-checks every claim. The best response wins.

4. Follow up — Ask follow-up questions to dig deeper. All models retain conversation context.

Core Features

ThunderScore
AI responses scored on 9 weighted dimensions: factuality (22%), anti-hallucination (15%), relevance (18%), depth (13%), clarity (10%), originality (7%), actionability (6%), conciseness (5%), confidence (4%).
🛡️
TruthLock
Automated fact-checking that verifies claims across models, detects hallucinations, flags fabricated citations, and cross-references responses for conflicts.
📊
BattleRank
Personal analytics dashboard tracking win rates, factuality scores, hallucination rates, and model performance trends across all your battles.
🔬
Web Verification
For factual queries, TruthLock searches the web to verify claims against real-world data, catching outdated or incorrect information.
🎯
Consensus Engine
Identifies what all models agree on and where they diverge, giving you a meta-view of AI reliability on any topic.
🔄
Follow-up Threads
Continue conversations with all models simultaneously. Each model retains full context from prior turns.

API Reference

POST /api/v1/battle

Run a battle programmatically. Requires an API key (Pro tier or above).

REQUEST
curl -X POST https://battledome.ai/api/v1/battle \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "query": "What is quantum computing?",
    "modelIds": ["claude", "gpt", "gemini", "grok"]
  }'
RESPONSE
{
  "winnerId": "claude",
  "results": [
    {
      "modelId": "claude",
      "thunderScore": 9.2,
      "trustLevel": "verified",
      "text": "..."
    }
  ],
  "evaluation": {
    "consensus": "...",
    "divergence": "...",
    "verdict": "..."
  }
}

Plans & Limits

Free
$0
3 battles/day
4 models per battle
1 free TruthLock demo
Standard models
10 battle history
Pro
$15/mo
50 battles/day
4 models per battle
TruthLock every battle
Web verification
Full BattleRank
Unlimited history
Team
$49/mo
Unlimited battles
4 models per battle
Everything in Pro
Premium models
API access
Priority support

FAQ

How does scoring work?
ThunderScore evaluates AI responses on 9 weighted dimensions. The model with the highest combined score wins. Scores range from 1.0 to 10.0.
What is TruthLock?
TruthLock is our fact-checking engine. It verifies claims, detects hallucinations, cross-references models against each other, and optionally searches the web for real-world verification.
Can I use my own API keys?
Yes! Go to Settings (⚙️) and add your API keys for any model. This unlocks premium model versions and doesn't count against your daily limit.
Is my data private?
Yes. Your battles are stored securely in your account. We never share your queries or results. See our Privacy Policy for details.
How do follow-up conversations work?
After a battle, use the follow-up input to ask another question. All models retain context from prior turns, allowing for deep multi-turn analysis.