Perplexity has officially restructured its collaborative research hubs, transforming its former “Spaces” feature into a new capability called “Projects” designed to inject deeper, more persistent context into AI-driven workflows. The update, rolled out to give users tighter control over their research environments, lets individuals and teams group documents, custom prompts, and specific instructions into dedicated workspaces.
According to product updates from the San Francisco-based conversational search startup, the shift addresses a growing demand for sustained context in automated research. While Spaces previously functioned largely as ad-hoc collections of threads, the new Projects structure aims to function more like a centralized operating folder, anchoring AI models to a curated set of files and rules across multiple sessions.
The transformation arrives as generative search platforms compete aggressively for professional and enterprise users who require rigorous fact-checking and document-driven answers. By allowing users to upload specific reference materials—such as financial reports, academic papers, or internal codebases—into a single Project, Perplexity attempts to reduce hallucinations and streamline specialized inquiries.
Understanding Perplexity Projects and Persistent Context
The core mechanic behind Perplexity Projects is persistent context management. In standard AI chat interfaces, each new query risks losing the nuanced constraints of previous conversations unless laboriously re-explained. Projects solves this friction by maintaining a persistent background layer of reference data.
Users can upload PDFs, text files, and spreadsheets directly into a Project, while also writing custom instructions that dictate how the AI should respond. For instance, a financial analyst can build a Project dedicated to a specific quarterly earnings cycle, uploading SEC filings and setting a directive that all answers must cite exact page numbers and table data. Subsequent prompts entered within that Project automatically draw from those uploaded parameters.
Industry analysts note that this approach mirrors workspace features introduced by rival conversational platforms, signaling a broader industry shift toward task-specific, file-grounded generative tools. Rather than acting as an open-ended oracle, the AI is constrained to act as an expert reader of a curated library.
Features and Workflow Integration
The transition from Spaces to Projects brings several concrete interface adjustments designed to improve organization. Users who previously relied on Spaces will find their existing collections migrated or adaptable to the new framework, though with enhanced file-handling capabilities.
- Custom Instructions: Users can define specific personas, output formats, or analytical lenses that persist across every chat thread initiated within the Project.
- Document Curation: Uploaded files serve as the primary knowledge boundary, allowing the search engine to prioritize internal documents over general web results when requested.
- Collaborative Sharing: Teams can share Projects to standardize research workflows, ensuring multiple contributors query the same verified dataset.
These controls are particularly useful for technical writers, legal researchers, and software developers who manage complex documentation sets. By keeping project files separate from everyday web searches, users avoid contaminating their specialized threads with unrelated queries.
Market Positioning and Enterprise Adoption
The rollout of Projects underscores Perplexity’s broader push into enterprise and professional tiers. As businesses grow increasingly cautious about data privacy and general-purpose AI inaccuracies, structured environments like Projects offer a safer bridge between large language models and proprietary information.
Competitors across the generative software landscape have rushed to deploy similar file-grounding features, emphasizing citation integrity and data security. Perplexity’s iteration relies heavily on its signature web-search grounding combined with localized file retrieval, attempting to bridge real-time internet data with private document archives.
Technical documentation and official support portals detailing the migration from Spaces to Projects are available directly through Perplexity’s platform updates, providing step-by-step guidance for existing subscribers looking to adapt their saved threads to the new architecture.
As AI developers continue refining context windows and retrieval-augmented generation techniques, the success of features like Projects will likely depend on how seamlessly users can manage large volumes of reference material without sacrificing speed. Further updates regarding enterprise-specific security controls and API integrations for Projects are expected as the feature scales globally.
What are your thoughts on transitioning from Spaces to Projects? Share your workflow experiences and join the discussion in the comments below.
Related reading