The Perplexity Blog

Isometric line drawing of cubes arranged on a grid, with a central column of cubes descending through stacked square frames.
Research1 Sept 2026Optimising On-Device Inference for Apple SiliconA custom local engine that improves prefill and decode throughput.
A floating, silvery metallic Mac mini against a black background, featuring intertwined loops and a small grid of glowing blue squares.
News1 Sept 2026Introducing Hybrid Compute on MacPerplexity Computer uses cloud AI and local models on Mac, keeping sensitive files and information on the device.

Featured

  1. News30 Jul 2026Spaces are now ProjectsProjects are a single place to manage, create, and collaborate on tasks with shared files and memory.
  2. Image of a globe against a dark background with nodes on it.
    News18 Jun 2026Self-improving memory for agentsBrain is Perplexity's self-improving memory system for Computer that learns from past work to deliver faster, more accurate results over time.
  3. Developers6 May 2026Introducing finance search in the Agent APIIn a single tool call, Finance Search combines licensed datasets, real-time market data, and cited sources.

More from the Perplexity Research team

  1. PII-TRACE: Detecting Personal Data Before It Leaves the DeviceA 13-language benchmark for consistent PII detection across long-running conversations, paired with a compact 0.6B detector designed to run locally.
  2. Brain: Agentic Memory as a Knowledge WikiA structured, traceable, self-improving Markdown filesystem, compiled offline and navigated on demand.
  3. Securing agents across Perplexity’s client endpoints with NumbatNumbat is Perplexity’s open-source agent security suite for client endpoints. It detects, prevents, and investigates risky AI agent behaviour on macOS, Linux, and Windows.
All research

Recently published

162 posts found

  1. AI integration: Getting ROI from your AI investmentConnecting the right data, tools, and workflows, AI is helping bring businesses real value by accelerating development, improving decisions, and automating repetitive work.
  2. Fast Embeddings on GPUsFast and accurate search is vital to all of Perplexity, from Search and Computer to our API Platform. Behind the scenes, the heavy lifting is done by embedding and ranking models, which help our systems identify the most relevant results for a given query. We achieve state-of-the
  3. PII-TRACE: Detecting Personal Data Before It Leaves the DeviceA 13-language benchmark for consistent PII detection across long-running conversations, paired with a compact 0.6B detector designed to run locally.
  4. AI personal assistants: What digital helpers can and can’t doHere’s what AI personal assistants are capable of, where they fall short, and how to use them well.
  5. How to Spot AI Hallucinations: 7 Red FlagsLearn seven signs of AI hallucinations, why models invent facts, and a five-minute checklist for verifying citations, claims, dates, and sources.
  6. Introducing Hybrid Compute on MacPerplexity Computer uses cloud AI and local models on Mac, keeping sensitive files and information on the device.
  7. Optimising On-Device Inference for Apple SiliconA custom local engine that improves prefill and decode throughput.
  8. Computer Connects to 20+ New Licensed Finance Data SourcesPerplexity Computer connects to new licensed finance data sources, including D&B, Guidepoint, IBISWorld, and 20+ others.
  9. A Local-First Agent for Private and Cost-Effective Knowledge WorkA harness and model co-designed for local knowledge work, running on device and accessing remote capabilities on demand.
  10. Introducing Portable Computer for local-first AIPortable Computer runs Perplexity Computer entirely on device with NVIDIA, keeping private data local and escalating to the cloud only when a task needs it.