Perplexity vs Claude: search engine or thinking partner

This is the rare comparison where the two products barely overlap. Perplexity is an answer engine: live web search with citations, wrapped in a model picker. Claude is a thinking partner: the strongest long-form reasoning with the best memory controls. Many people rightly use both. We build NodePad, so read the third column knowingly: it competes with neither core loop, no live search and no house model, but it is where the work between the answers happens. Facts from Perplexity, careful drafts from Claude-class models, and every thread of the project visible on one canvas that hands any model exactly the context you chose. Quick cited answers: Perplexity. One deep thread: Claude. A project that outgrows both shapes: that is the third column.

dimension Perplexity Claude NodePad
Core loop Ask, get a cited answer from live web search, follow up. A long conversation with a careful model; artifacts for the outputs. A canvas of conversations about one problem; fork, reference, merge.
Price of entry Free; Pro $20/mo; Max $200/mo. Free; Pro $20/mo ($17 annual); Max $100 to $200/mo. Free on community models; paid from $5/mo.
Models you can use The broadest picker in class: GPT, Claude, Gemini, Grok, and Sonar, but always inside Perplexity's search pipeline. Anthropic models only (Haiku, Sonnet, Opus, Fable). Claude, GPT, and Gemini, each receiving the bare context you assembled; no pipeline in between.
Who assembles the context Perplexity does; retrieved snippets and its own prompting surround your query invisibly. Memory synthesis and chat search decide what enters; editable summary, not per-message control. You. The path you built plus what you referenced or merged. Nothing hidden, nothing injected.
Branching a conversation None; threads are strictly linear, grouped into Projects. Edit and regenerate keep hidden versions behind per-message arrows; no view of the tree. Fork any point onto a visible parallel path; the original continues untouched.
Combining threads No merge; exploring two hypotheses means two threads with duplicated context. No cross-chat merge; Projects share knowledge, not conversations. Merge several branches or chats into scope and ask for the synthesis.
Research The differentiator; live search, Deep Research reports, and Model Council on Max. Research mode from Pro up; strong, but not a live answer engine. No built-in retrieval; plugins can add tools. Perplexity wins this row outright.
Usage limits Deliberately vague ("a high volume of daily searches"); quotas shift without notice. A conversation budget with no published numbers. A visible meter; add usage when it runs out.
Try NodePad free

No credit card to start. Every paid plan reaches every model.