Skip to main content
Vincony
AI OSPricingTrust
Log inStart Free
Free Credits
  1. Resources
  2. Ai Search Guide
Back to blog
Tutorials

AI Search Guide: Web Research with Citations, From Quick Answers to Deep Reports

Vincony TeamJune 6, 20268 min read

# AI Search Guide: Web Research with Citations, From Quick Answers to Deep Reports

Whether you need a fast answer to a topical question or a structured deep-dive report you can export to Google Docs, Vincony's AI Search hub has you covered. This guide walks through both panels — Quick for single-shot web lookups and Conversational for iterative, citation-rich research — so you can get maximum value from every credit you spend.

---

What Is AI Search?

AI Search is Vincony's consolidated web-research surface. It replaced the older Search Agent and Deep Research tools (both now redirect here) and lives at /os/ai-search. The hub exposes two depth modes selectable via the tab bar at the top of the page:

ModeTabBest forCredit cost
QuickQuick (?panel=quick)Single-question lookups, fact checks, news1–3 cr per search
ConversationalConversational (?panel=conversational)Multi-step research, long-form reports, follow-ups3 cr per search + 1 cr per follow-up

Both modes use live web data. Neither relies on a training-data snapshot — the answer you get reflects what is on the web right now.

---

Panel 1 — Quick Search (Fast, Cited Single Answers)

The Quick panel is designed for speed. Type a question, hit Enter, and get a synthesised answer in seconds with numbered citations you can click through to the source.

How it works under the hood

Quick search defaults to Perplexity Sonar mode. Short queries (ten words or fewer) cost 1 credit and run Sonar; longer or more complex queries are automatically upgraded to Sonar Pro at 3 credits for a higher-quality synthesis. You can also toggle to Firecrawl mode (the switch labelled with a lightning bolt) which scrapes and summarises up to eight live pages instead — useful if you want raw page content alongside the summary.

Step-by-step: your first Quick search

  • Open AI Search and click the Quick tab (or append ?panel=quick to the URL).
  • Type your question in the search box. Example: "What are the best open-source LLMs in 2026?"
  • Press Enter or click Search. The credit cost is shown in the button badge before you commit.
  • Watch the AI Summary card populate in real time as the answer streams in.
  • Below the summary, a Sources section lists every cited URL. Click View content on any source card to expand the scraped markdown inline.
  • Once the answer arrives, use the Follow up field (visible in Sonar mode) to refine your question — each follow-up costs 1 credit and is answered in the context of what was just retrieved.
  • Hit Copy to copy the plain summary, or open Export and Share to save directly to Google Drive, Dropbox, Notion, Google Docs, or OneDrive — or send by email.

Quick search tips

  • Toggle Sonar on for most queries. It is faster and cheaper than Firecrawl for factual Q&A.
  • Use the time filter (visible in Firecrawl mode) to restrict results to the past hour, day, week, month, or year when you need recency.
  • Recent search chips appear below the input if you have typed queries before. Click any chip to instantly re-run that search.
  • Free plan users get a limited monthly credit allocation. Starter and above get significantly more — see pricing.

---

Panel 2 — Conversational / Deep Research (Multi-Step, Structured Reports)

The Conversational panel is where Quick ends and deep work begins. It is powered by Perplexity Sonar Pro on every query, which means richer source coverage, inline citations, and auto-generated follow-up question suggestions after every answer.

How it works under the hood

Each search deducts 3 credits and streams the response token by token so you see progress immediately. Vincony strips the raw follow-up section from the stored answer and surfaces those suggestions as clickable buttons beneath the result card — tap one and the next search launches automatically, optionally carrying forward context from your prior session if Context on mode is active.

Answers are saved locally (up to ten) so your research history persists between sessions. You can pin any result card (the pin icon in the top-right of the card) to keep it accessible above your rolling history, which is handy when you are cross-referencing multiple sub-topics.

Step-by-step: running a deep-research session

  • Open AI Search. The Conversational tab is the default — you will land here automatically.
  • Type your opening research question. For complex topics, start broad: "What are the main regulatory frameworks governing AI in the EU in 2026?"
  • Press Enter or click Search (3 credits). The answer streams in and a list of numbered citations appears below when complete.
  • Scan the auto-generated Follow-up questions chips at the bottom of the result card. Click one to drill down — for example "How does the EU AI Act define high-risk systems?" The new query runs with context from the previous answer if Context on is active.
  • To switch topics cleanly, click New search in the toolbar — this resets context so earlier answers do not bleed into unrelated questions.
  • For academic work, toggle the Academic button in the filter toolbar. Sonar Pro shifts its source weighting toward scholarly and peer-reviewed content.
  • Need to pull from specific sites? Click Domain filter, enter comma-separated domains (example: arxiv.org, nature.com) and optionally prefix a domain with a minus sign to exclude it.
  • When you are satisfied, open Save / Share on any result card to export to Google Drive, Google Docs, or OneDrive, or send by email via your connected account.
  • Pin your most important cards before you clear history. Pinned cards persist until you explicitly unpin them.

Conversational mode tips

  • Context on vs. Context off. With context on, the last three query-answer pairs are injected into each new request. This is powerful for iterative refinement but can cost you accuracy if you pivot to a completely different topic — hit New search to start fresh.
  • Stop mid-stream. If the answer is going in the wrong direction, click the red stop button that appears next to the search button while streaming. If no content has arrived yet, your credits are refunded automatically.
  • Keyword shortcut. Press / from anywhere on the page to jump focus to the search input. Press Escape to clear it.
  • Retry on failure. If a search fails (network error, model timeout), a Retry button appears on the error card and re-runs the same query without requiring you to retype it. Credits are refunded on failure.
  • Word count and timestamp. Every result card shows the model used (Perplexity Sonar Pro), the word count of the answer, and when it was retrieved — useful for knowing how fresh the information is.

---

Choosing the Right Mode

Use Quick when you want a fast, cite-backed answer to a single question and do not need to iterate. Use Conversational when you are building up an understanding across several related sub-questions, want richer sourcing, or need an answer long enough to be worth exporting as a document.

Both modes sit inside the same AI Search page — switch between them freely at any point. Your Quick history and Conversational history are stored separately in browser storage.

---

Plans and Credits

Stay ahead in AI

Get our weekly AI insights — tips, model comparisons, and guides delivered to your inbox.

No spam, unsubscribe anytime.

AI Search is available on all plans, but credit budgets differ:

  • Free — enough credits to try both modes, limited monthly quota.
  • Starter — suitable for light daily research.
  • Pro — comfortable for multi-step research sessions across several topics per day.
  • Power and Business — designed for heavy usage, team workflows, and integrations with Notion, Drive, and email.

See the pricing page for exact credit allocations per plan.

---

What to Try Next

Once you have a research answer in hand, Vincony has tools to act on it:

  • Cross-check any claim with the Fact Checker for structured verification with confidence scores.
  • Save key findings to Second Brain to build a persistent knowledge base you can query later.
  • Curious which AI model powered your answer? Visit Models for a full breakdown of what is available at each tier.

Start with a question you have been meaning to research. Type it into AI Search and let the citations do the heavy lifting.

Actions

Related Articles

Tutorials

How to Build a Custom AI Chatbot for Your Website in Under an Hour

Feb 13, 2026

Tutorials

AI Image Generation Tips & Tricks: From Prompt to Pixel-Perfect Output

Feb 20, 2026

Tutorials

AI Video Generation Tutorial: Create Professional Videos from Text Prompts

Feb 21, 2026

On this page
Vincony

Access the world's most powerful AI models through a single, unified platform.

Product

  • All Models
  • Chat
  • Image Generation
  • Video Generation
  • Voice Studio
  • Song Studio
  • All Tools
  • Pricing
  • Integrations
  • API & Developers
  • Download Apps

Solutions

  • Use Cases
  • By Role & Industry
  • Case Studies
  • Testimonials
  • Marketplace
  • Templates
  • Agency Portal
  • White-Label

Resources

  • Help Center
  • Guides
  • Glossary
  • Blog
  • Changelog
  • Feedback
  • Savings Calculator
  • Credits Calculator
  • Plan Recommender

Company

  • About
  • Contact
  • Contact Sales
  • Security
  • Trust Center
  • Bug Bounty
  • System Status
  • Partners
  • Affiliate Program
  • Refer & Earn
  • Brand & Media

Legal

  • Terms of Service
  • Privacy Policy
  • Data Processing Agreement
  • Acceptable Use
  • Cookie Policy
  • Refund & Cancellation
  • Accessibility
  • Sub-processors
  • DMCA & Copyright
Compare AI platforms·Best AI tools·All alternatives·Sitemap

© 2026 VINCONY AI LTD (17047337). All rights reserved.

VINCONY AI LTD · Company No. 17047337 · 3rd Floor, 86-90 Paul Street, London EC2A 4NE, England

GDPR Ready · CCPA Compliant · SOC 2 Aligned · 256-bit Encryption ·

Get weekly AI tips & updates