Lens

Lens is a local AI research agent that takes a query, performs structured research, and stores results in a persistent workspace.

Installation

pip install lensdev

Usage

Run a research task:

lens research -q "how do vector databases work"

Workspace System

Every run generates a workspace:

workspace/<session_id>/

Example:

workspace/session_20260604_143210/

Structure:

• report.md (Main research report)
• sources.json (Combined raw data from all collectors)
• meta.json (Session metadata: query, timestamp, version)
• raw/ (Individual collector results)

CLI Commands

Run research:

lens research -q "query"

Output:

Lens.
────────────────────────────────────────────────────────────────────────────────

Research
your query

Collecting Sources
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% (7/7)

Research complete.
Workspace saved at: workspace/session_20260604_143210/

List sessions:

lens list

Output:

 Session ID                        
 session_20260604_143210
 session_20260604_104908

Resume session:

lens resume <session_id>

Output:

Resuming session: session_20260604_143210
Query: how do vector databases work
Path: /path/to/workspace/session_20260604_143210
Re-generating reports...
Reports re-generated.

Output Behavior

After completion:

Research complete.
Workspace saved at: workspace/<session_id>/

Philosophy

Lens turns research into structured, persistent workspaces instead of temporary terminal output.

GitHub: https://github.com/developer8sarthak/lens-research
PyPI: https://pypi.org/project/lensdev/
Documentation: https://lensdev.pages.dev/