Perplexica – v1.12.0
🚀 Perplexica v1.12.0 just dropped—and it’s a full rebuild, not just an update!
✨ No more Langchain chaos—custom streaming engine now gives you precise control over response flow, with smooth partial streaming & function calling.
💬 Sessions remember you—refresh? Disconnected? It auto-resumes and lets you replay streams.
🤖 New DIR agents + live widgets (weather, stocks, math) run in parallel while you research.
📚 Deep Research Mode (Beta): LLM thinks step-by-step, scrapes, discusses, uploads—builds full reports like a relentless research assistant.
📄 Smart File Search: Upload huge PDFs/DOCX, search across them all with RRF reranking.
🎯 Sources over modes: Toggle web/academic/discussions—or go fully offline with your files.
🎨 UI glow-up: Framer Motion animations, syntax-highlighted code blocks, theme fixes, version info in settings.
🧠 Query classification + action system: Every query gets classified first—tool calls now drive research (Google Workspace? Coming soon).
🧩 Block & JSON streaming: Watch every step of the LLM’s thinking in real time. No more “is it stuck?” panic.
You’re not searching anymore—you’re investigating.
Try Deep Research Mode with a complex topic. You won’t believe the depth. 🧠💥