Perplexity

Today we're launching Projects, an evolution of Spaces. Projects are hubs for ongoing work in Computer. A single place to manage, create, and collaborate on tasks with a shared file system and persistent memory. Each Project has a shared file system. Computer reads, edits, and saves Project files across sessions, so the work builds on itself with each task. When collaborating with teammates in a shared Project, everyone works from the same source of truth. Projects connect to Brain, Computer's self-improving memory system. Between tasks, Brain reviews the Project's files and sessions and updates what it knows, so each task starts with full context from previous work. Projects are available today to all users. Read more: https://lnkd.in/gPRgtkr3

Today we’re open-sourcing Numbat, an agent-detection and response layer that is designed to work across agent harnesses. It gives security teams visibility into agent activity, with controls to block selected actions before execution. Numbat works across agent harnesses on desktop, CLI, IDE, and gateway agents. It supports live monitoring, local detection, pre-action blocking, and forensic reconstruction from existing session artifacts. Numbat runs across our systems and sends audit events, findings, and security alerts to Computer. Computer continuously analyzes this telemetry, escalates suspicious behavior, and proposes new on-device detections for Numbat, creating a continuously improving detection flywheel. Numbat is available as a single Go binary for macOS, Linux, and Windows. Read more: https://lnkd.in/gHuVv3uu Available under Apache 2.0: https://lnkd.in/g-XbWrUw

We've added Kimi K3 to Perplexity and Perplexity Computer for Pro and Max subscribers. Kimi K3 in Perplexity is hosted exclusively on U.S.-based servers.

Model Council is now available inside Computer. Run independent analysis across multiple frontier models, choose your analysis depth, and get a single cited report on where models agree, where they disagree, and what each found that others missed. Read more about using Model Council inside Perplexity Computer: https://lnkd.in/gsZbf_uE

Personal Computer is now available in the Perplexity app for Windows. Personal Computer is the local agent harness for your work. It orchestrates agents across your local files, connected apps, and the web. Research, code, browse, and build inside one unified platform. Personal Computer routes work to subagents using 15+ models. It can open, read, and edit files inside approved local folders on your desktop. Every task starts with full context from your past tasks, files, and projects through Brain, our persistent memory system. It pulls data from 400+ connected apps into your local files, and works with your existing Microsoft Excel, PowerPoint, Word, and Outlook files. Available now for Pro, Max, and Enterprise subscribers on Windows 10 and 11. Download: https://lnkd.in/e5tnVq98 More on connecting Microsoft 365 apps to Computer: https://lnkd.in/esRW_sak

We're hosting a one hour conversation with Perplexity CISO, Kyle Polley, on August 12. The session covers how security is built into Computer, recent security releases like SPACE, and what using Computer looks like day to day for a CISO. Register below:

Claude Opus 5 is now available in Perplexity and Perplexity Computer. We evaluated it against six other models on WANDR. It outperformed all but Fable 5, while being 57% cheaper.

We're hosting a live workshop on how lawyers are using Perplexity Computer across different practices of law while staying firmly in control of their output. Join Alison Grigonis, our Senior Director of Product, Privacy, and Regulatory Legal and Jenny S., our Product Marketing Lead on August 4 at 12pm ET. Register below:

Introducing SPACE, the sandbox platform behind Perplexity Computer. It creates isolated environments for code, files, and long-running agent sessions. SPACE has handled 100% of Computer production traffic since June. https://lnkd.in/gxNTgx-Q Agent infrastructure must be functional, efficient, and secure and traditional sandboxes were built for short-lived code execution. Agents need to run code, edit files, and run for hours or days. Runtimes must preserve work without leaving credentials inside environments. SPACE separates the session from the sandbox running it. Each task gets a disposable Firecracker microVM that is destroyed when the work ends. Rolling snapshots preserve live memory and files, so the session can pause, resume, or branch across sandboxes. On identical production traffic, SPACE reduced median sandbox creation latency from 185 ms to 60 ms. P90 fell from 447 ms to 89 ms. Last week it handled millions of sandbox creations and tens of millions of reconnects for Computer. Read more: https://lnkd.in/gSFrn7bW

We’re open sourcing WANDR. WANDR is an internal benchmark we built and used for building deep and wide research capabilities inside Perplexity Computer. https://lnkd.in/gfCh95F3 Wide-and-deep research requires two capabilities. 1) Agents must search broadly enough to find all qualifying entities 2) Agents must investigate deep enough to support every claim with evidence. WANDR represents these requirements as hierarchical, independently verifiable records. It consists of 500 research tasks that require 170,495 source-backed records across three tiers of difficulty. WANDR tests how well agents discover large sets of entities and verify specific facts about each one. It provides a dense, interpretable eval signal that reveals whether an agent fails, and where. The pipeline also doubles as a semi-automated factory for training data. WANDR is constructed from de-identified production use cases covering people’s day-to-day research tasks like competitive research, due diligence, literature review, market analysis, product comparison, talent sourcing, and more. Instead of grading against a gold solution, WANDR re-fetches every cited page and checks each claim against the underlying evidence. This allows certain tasks to contain time-varying facts. Soft scores give partial credit. Hard scores require every component for a member to be complete and correct. The benchmark tasks and evaluation harness are available at: https://lnkd.in/gvaYBDdG