🚀 What's New in Perplexity: Game-Changing Search API + Major Updates
Perplexity just dropped their next-gen Search API, and it's a kinda game-changer for anyone building AI apps, automations, or research workflows.
Besides the API, one Unofficial Feature first:
🎥It can also now create videos on its own. Excellent. They're comparable to VEO 3 in quality.
They don't contain sound, though. BUT, if you ask perplexity, it will deny that it can create videos.
Which is correct, because it's not yet aware it can. LOL.
🔥 The Search API: What Makes It Different
Instead of just returning a bunch of links, this API gives you clean, AI-ready answers with citations, metadata, and structured output. Here's what you can do with it:
Four Powerful Search Modes:
⚡ Instant Search
- Lightning-fast results (basically instant)
- Returns titles, URLs, snippets, dates
- Perfect for quick lookups and lightweight automations
- Comparable to Google Search API or Exa
📚 Grounded Search
- Uses AI models (sonar, sonar pro, sonar pro reasoning, deep research)
- Every answer is backed by verified citations
- Takes longer but gives you rock-solid, traceable sources
- Ideal for research and content that needs credibility
🎓 Academic Search
- Filter by specific domains (archive.org, PubMed, etc.)
- Set recency filters (past month/year)
- Prioritizes peer-reviewed studies
- Takes 7-10 seconds but returns the highest-quality academic sources
🎯 Structured Output
- Define your own JSON schema
- Get responses in exactly the format your app needs
- Perfect for vibe-coding apps in minutes
- No more wrestling with inconsistent data structures
🤖 Multi-Agent Workflows
Build your own AI research team:
- Set up a main agent that delegates to specialized sub-agents
- Each sub-agent runs different search types (academic, practical, innovative)
- Main agent consolidates everything into one comprehensive answer
- Identifies patterns, contradictions, and recommendations automatically
💰 Pricing & Performance
- ~$5 per 1,000 requests (varies by model)
- Latency: Instant ≈ real-time Grounded ≈ a few seconds Academic ≈ 7-10 seconds
🆕 Other Major Perplexity Updates
API & Developer Tools:
- API Key Rotation for better security
- Standalone Search API (raw ranked results without AI processing)
- API Portal for team organization and usage monitoring
- Advanced Sonar Models with improved reasoning
User-Facing Features:
- iOS App Redesign: Faster queries, gesture navigation, smoother UI
- Multimodal Search: Upload images to search (now available for all users)
- Location-based Filtering: Filter results by your location
- Open-source Tools: Evaluation frameworks to benchmark search quality
💡 Real-World Use Cases
Build in Minutes:
- Academic research portals with custom filters
- Multi-agent research assistants
- Automated content pipelines with structured data
- Custom search engines with source verification
Integration-Friendly:
- Works with n8n, Claude Code, Cursor, Windsurf
- Just plug in curl requests from their docs
- Copy/paste templates and swap in your API key
🎯 Bottom Line
This API lets you build custom, source-backed, structured search workflows without heavy engineering. Whether you're automating research, building AI agents, or vibe-coding apps, Perplexity just made it 10x easier.