Perplexity upgrades Mac app with local models
Perplexity updated its macOS desktop application with faster responsiveness, native preview search, and real-time task redirection via a new "Send now" control. The client pairs local Apple Silicon execution with cloud frontier models to enable private, hybrid compute workflows.
Perplexity's rapid desktop iteration proves that winning daily AI workflows requires deep OS integration and on-device compute rather than another browser tab.
- –Hybrid compute routes sensitive files and local context to on-device models like Qwen or PPLX Pro while delegating heavy reasoning to cloud frontier LLMs.
- –Live task redirection via "Send now" addresses a major UX friction in multi-agent workflows by letting users steer running agents without restarting.
- –Native macOS ergonomics like ⌘F across Markdown tables and ⌘⇧T tab restoration bring agentic workspaces closer to standard developer tooling.
- –Providing fully local offline execution on Apple Silicon offers a privacy guarantee essential for enterprise and proprietary codebases.
DISCOVERED
16m ago
2026-11-20
PUBLISHED
51m ago
2026-11-20
RELEVANCE
AUTHOR
AravSrinivas