Perplexity introduced Labs for longer projects that could return files, dashboards and simple web apps
Perplexity introduced Labs on May 29, 2025 for Pro users, combining research, code execution and asset creation into longer project runs.
Historical reporting: Source 0 is Perplexity’s dated Labs announcement. Source 1 is a separate dated Perplexity changelog record for the same release day. Both describe the 2025 Labs surface; neither establishes later product naming, limits, retained files or current subscription terms.
Key takeaways
- Perplexity introduced Labs on May 29, 2025 as a Pro-user mode for projects such as reports, spreadsheets, dashboards and simple web apps that could take ten minutes or more.
- The release described a tool set spanning deep web browsing, code execution, chart and image creation, plus an Assets tab for files generated during a project.
- A generated dashboard or application remains an artifact to inspect: its source data, formulas, code, permissions, accessibility and deployment behavior need separate verification.
A research prompt could become a multi-artifact project[source 1][source 2]
Perplexity introduced Labs on May 29, 2025 as a new mode for work that extended beyond a single research response. The company said Labs could make reports, spreadsheets, dashboards and simple web apps, and that a task could take ten minutes or more to complete. That duration matters because it signals an attempted project workflow: the system was presented as gathering information, analyzing it and turning results into files or an interactive output. The announcement described availability for Pro subscribers. It did not say that every prompt would produce a correct deliverable, that a generated app could safely access private systems, or that a result met the quality standards of a spreadsheet owner, designer, engineer or subject-matter reviewer.
The stated tool set joined retrieval, computation and presentation[source 1]
The launch named deep web browsing, code execution, chart creation and image creation as capabilities used during Labs work. Perplexity also described an Assets tab where users could inspect and download generated material such as charts, images, code files and other artifacts. This is a more concrete claim than a generic promise that an assistant can help with projects: it describes several kinds of intermediate and final output. It also exposes several independent places where a result can fail. A chart can use the wrong denominator, a spreadsheet can contain an opaque formula, an image can make a misleading claim, and generated code can have missing tests or unsafe dependencies. Downloadability is useful for review, but it does not make the file a verified source of truth.
Pro availability and examples framed a product lane, not a guarantee[source 1][source 2]
Perplexity’s launch examples included building trading strategies, dashboards, headless browsing tasks for real-estate research, mini web apps and storyboards. The contemporaneous changelog described Labs as available to all Pro users and offered broad work and personal examples, including a marketing campaign, finance analysis and meal planning. These examples show the intended range of prompts, but they do not prove that every underlying website was accessible, that a financial output was investment advice, or that a web app had a security review. The relevant historical fact is the supplier’s defined product offer on May 29. Eligibility, output quality, data rights and any cost implications must stay tied to the source’s stated terms and the user’s own account context.
Longer autonomous work still needed owner review[source 1][source 2]
Labs represented a shift from answer generation toward a project bundle with research, computation and presentation steps. That can help a user get to a reviewable starting point sooner, especially when a task needs multiple formats. It does not remove the obligations that come with each format. A team should review data provenance before publishing a dashboard, test generated code before exposing it, check licenses and permissions before distributing assets, and confirm that a reported conclusion follows from its inputs. The release did not provide an independent evaluation of accuracy, reliability or safety across the named use cases. It is therefore sound to treat the Labs launch as product history and the files it could produce as candidate artifacts, not as automatic approval for action.
Sources
Continue reading
Related reading
News analysisPerplexity opened Deep Research to free and Pro users as a report-generating search mode3 min read
News analysisAnthropic donated Agent Skills to AAIF: treat open governance as a new supply-chain input3 min read
News analysisZoom released AI Companion 2.0 as a cross-Workplace context and action surface3 min read