OmniQra Documentation

Docs & Guides

Everything you need to get the most out of OmniQra — from your first question to advanced workflows. 📚✨

🚀

Getting Started

OmniQra takes about 30 seconds to set up. No credit card, no install, no waitlist.

Go to omniqra.com
Open the site in any modern browser — Chrome, Safari, Firefox, Edge or Brave all work.
Sign up free
Use Google sign-in or email. You're instantly granted 3 free credits per day — no card required.
Ask a question
Type your prompt in the chat box and hit Enter. Multiple AI models start answering in real-time.
Read the synthesis
After both candidates finish, OmniQra's green synthesis box gives you the final, cross-checked answer.
👤

Creating an Account

An account lets us track your credit balance, save your chat history, and keep your conversations private.

Sign-in options

  • Google: one-click sign-in, no password to remember.
  • Email + password: standard sign-up flow with email verification.
Both methods give you the same daily free credits. Choose whichever feels faster.

Forgot your password?

Hit "Forgot password" on the login screen. We'll email a reset link valid for 1 hour.

💬

Your First Question

OmniQra works best when you ask focused, specific questions. Here are some great starter prompts:

  • "Compare React Server Components and Next.js App Router in 5 bullet points."
  • "Write a polite email asking a client for delayed payment."
  • "Explain quantum entanglement like I'm 12, with one analogy."
  • "Plan a 7-day Tokyo itinerary for a first-time visitor in spring."
💡
Pro tip: the more context you give, the better all 3 models perform. "Audience, format, length, constraints" works wonders.
🧠

How OmniQra Works

Every question you send goes through a 5-stage pipeline in ~4 seconds:

1. CAPTURE → prompt sanitized, length checked 2. DISPATCH → sent in parallel to 2 candidate models 3. STREAM → both candidates stream tokens live to your screen 4. SYNTHESIZE → 3rd model reads both, writes the unified answer 5. PERSIST → saved to your chat history, credit debited

For a deeper dive, check out the Science Behind OmniQra page.

🤖

The Models

OmniQra runs three top-tier LLMs per question:

  • GPT-5 family — OpenAI's flagship. Great at structured reasoning, code, and instructions. Appears as Candidate A (orange).
  • Gemini 2.5 Pro — Google's top model. Strong on breadth, edge cases, and long-context reasoning. Appears as Candidate B (red).
  • OmniQra Synthesizer — reads both candidates and writes the unified answer. Always in the green box.
⚠️
You don't pick the model. OmniQra always uses all three — that's the whole point. Single-model selection defeats the cross-check.
🧬

Reading the Synthesis

The green OmniQra box is the answer to trust. Here's how to read the full output:

  • Orange box (Candidate A) — raw answer from one model. Useful for comparing.
  • Red box (Candidate B) — raw answer from the second model.
  • Green box (Omniqra) — the final synthesized answer that cross-checks both.
🎯
When the two candidates strongly agree, the synthesis is very high confidence. When they disagree, the synthesis explicitly flags the uncertainty.
✍️

Prompting Tips

✅ Do

  • Give context — who's the audience, what's the goal?
  • Specify the format — "in 5 bullets", "as a table", "as JSON".
  • Set constraints — word count, tone, language.
  • Ask follow-ups in the same thread to build on previous context.

❌ Don't

  • Don't dump 5,000 words and ask "thoughts?" — be specific.
  • Don't rely on OmniQra for legal, medical or financial decisions without verification.
  • Don't ask for real-time data (live prices, scores) — models have a knowledge cutoff.
📏

Limits

  • Per-message limit: 6,000 characters.
  • Free daily limit: 3 questions per 24-hour rolling window.
  • Rate limit: max 1 question every 3 seconds (to prevent abuse).
  • Thread length: unlimited, but very long threads may truncate earliest messages from model context.
🗂️

Chat History

All your conversations are saved to your account, encrypted, and only visible to you.

  • Open the sidebar to browse, search and rename past threads.
  • Click the ⋯ menu on any thread to delete or export it.
  • Threads are private by default — no sharing unless you generate a share link.
💎

Credits & Plans

OmniQra runs on a simple credit system: 1 question = 1 credit.

Free tier

Every account gets 3 free credits per day. Resets every 24 hours. No card, no signup tricks.

Top-up credits

Need more? Buy credits as a one-time top-up. They never expire. Add credits →

Subscriptions

Power users can subscribe for a monthly bundle at better per-credit pricing. View plans →

⚠️
Credits are non-refundable once purchased. See our refund policy.
💳

Manage Billing

  • All payments processed securely by Stripe — we never see your card number.
  • Update card, invoices and tax info from Account → Billing.
  • Cancel any subscription any time. Access continues until the end of the paid period.
  • Receipts are auto-emailed after every payment.
🔐

Privacy & Data

  • Your prompts are never used to train AI models.
  • Chat history is encrypted and visible only to you.
  • You can delete any thread — or your entire account — any time from settings.
  • Fully GDPR compliant. Full details in our policies.
⌨️

Keyboard Shortcuts

  • Enter — send message
  • Shift + Enter — new line in prompt
  • Ctrl / + K — focus search / new chat
  • Ctrl / + B — toggle sidebar
  • Esc — cancel in-progress generation
🔧

Troubleshooting

A model is taking too long

Refresh the page. If one model is down, OmniQra automatically falls back to another so your question still completes.

"Out of credits" error

Your daily free credits reset after 24 hours, or you can top up instantly.

Login link not arriving

Check spam folder. Sender is no-reply@omniqra.com. If still missing after 5 minutes, contact support.

Streaming stops mid-answer

Usually a network blip. The full answer is still saved in your thread — refresh and it will appear.

Frequently Asked Questions

Is OmniQra free?
Yes — 3 free questions every day, forever. No card required. Power users can buy credits or subscribe.
Which AI models does OmniQra use?
Currently OpenAI's GPT-5 family and Google's Gemini 2.5 Pro as candidates, plus a dedicated synthesizer model. We rotate to the strongest models available.
Do you train AI on my prompts?
Never. Your prompts are used only to generate your answer, then stored privately in your chat history.
Can I use OmniQra commercially?
Yes. You own the prompts you submit and have a non-exclusive license to use the AI responses for any purpose, personal or commercial.
Why is the answer different from ChatGPT or Gemini?
Because it's synthesized from multiple models cross-checking each other. That's the whole point — fewer hallucinations, more balanced.
Can I export my chats?
Yes. Use the ⋯ menu on any thread to export as Markdown or JSON.
Does OmniQra support images, files or voice?
Text-only for now. Image and file uploads are on our roadmap.
Is there an API?
A public API is on the roadmap. Email api@omniqra.com if you'd like early access.
Can I get a refund?
Credits are non-refundable once purchased. See our refund policy for full details.
How do I delete my account?
Go to Settings → Account → Delete Account. All your data is erased within 30 days.
Still stuck? We're human, too 👋
Email us at support@omniqra.com — real replies within 24 hours.
Contact Support →