Platform layers
Reasoning and planning
Purna interprets the request, identifies required information and tools, and can prepare an approval-gated plan for complex work.
Knowledge and data
Literature, scientific databases, project context, uploaded files, and connected reference libraries provide evidence and working material.
Compute and biocompute
General code and data handling run alongside specialized biological computation such as structure prediction and omics analysis.
Outputs
Results can include cited answers, papers, scripts, tables, datasets, figures, molecular structures, and downloadable files.
Ways to work
Standard sessions
Ask a question, attach or mention files, and continue with follow-up requests in the same session. Use this for focused research questions, database retrieval, file inspection, and iterative analysis.Plan Mode
For multi-step work, Purna can clarify the request and prepare an approach, assumptions, constraints, validation steps, deliverables, and success criteria for approval. After approval, the task runs through the background execution layer.Deep Research
Deep Research performs a directed literature review with source screening, evidence extraction, gap identification, and a paper-writing handoff. The research process remains inspectable through its overview, timeline, plan, evidence, sources, and gaps.Bioinformatics
Upload biological data and describe the comparison, method, thresholds, and outputs you need. Purna can inspect the files, run reproducible analyses, and return scripts, quality-control outputs, statistical tables, figures, and an interpretation.Paper Writing
Use Write Paper to create and edit manuscripts, reports, literature reviews, and other scientific documents. Paper Writing can also receive completed work from Deep Research.Consilium
Consilium uses multiple expert perspectives to examine a scientific hypothesis, surface disagreements, and produce a structured synthesis.The session workspace
The session is the central unit of work. A session can contain:- The original request and follow-up messages
- An approved roadmap for planned work
- Notes associated with the session
- Uploaded and generated files
- Compute instances and background jobs
- Deep Research activity
- Open papers, scripts, figures, tables, or molecular structures
Capabilities and knowledge sources
The controls beside the message composer determine what Purna can use.- Capabilities enable actions such as code execution, scientific image generation, and specialized biological models.
- Skills and knowledge control available scientific sources, connected libraries, and source-aware guidance.
- Availability can vary by plan, organization configuration, and provider access.
Files and projects
Projects group sessions, files, papers, jobs, and results around a study or research direction. Files can be uploaded directly, attached to a message, or referenced with@.
Generated outputs are saved back to the session and Files area so they can be inspected, downloaded, or used in follow-up work.
Runtime and billing
A single request can use both general computation and specialized biological computation. Each operation is metered according to the runtime that performs it. See Usage and billing for the distinction between Compute Runtime and Biocompute Runtime.Research responsibility
Purna is a research and decision-support platform. Review source material, assumptions, generated code, statistical methods, model confidence, and final outputs before using them in publications or high-stakes scientific, clinical, regulatory, or experimental decisions.