Product Changelog: Rapid Development & Version Updates
Everything we've shipped—new capabilities, faster workflows, deeper intelligence. Updated weekly.
This is a living document. We ship fast and often—new capabilities, workflow improvements, and fixes that make jhana more useful for your practice. Check back weekly for the latest updates.
How to Read This
Major Releases are transformative features that change how you work. New Features add capabilities you asked for. Improvements make existing features faster or smarter. We don't list every bug fix—only the ones that meaningfully affect your workflow.
December 2025
Drafting Agent: AI Document Editing with Track Changes
The Drafting Agent brings full AI-powered document editing directly into your workflow—with formatting, track changes, and version control. This is the missing link between research and production-ready documents.
Handoff from Paralegal: seamlessly move from research to drafting in one flow
Split-screen Cursor UX: see your document and AI suggestions side by side
Full Word formatting: headings, paragraphs, lists, emphasis—all preserved
Track changes and commenting: see exactly what AI suggests before accepting
Manual takeover: switch between AI edit and manual edit modes instantly
Version control with revert: undo any change, compare document states
Edit existing Suit files or generate new documents from scratch
Smart routing: AI determines whether to draft new or edit existing documents
Tag constituent files in chat: @mention Suit files so the Drafter knows which document to edit
What this means for you: Stop copy-pasting between AI chat and Word. Research, draft, revise—all in one place, with full control over what goes into your final document.
Paralegal Persistence: Background Runs & Session Recovery
Long research sessions no longer require you to keep the tab open. Paralegal now runs in the background and picks up exactly where you left off.
Background execution: close your browser, research continues
Seamless reconnect: return to see progress and stream continuity
Session recovery: graceful handling of disconnects, no lost work
Explicit stop: research only stops when you hit the kill button
Multiple concurrent chats: run up to 5 parallel research sessions
Session-based connection: chats persist through renames
XL by jhana: Deep Research for Complex Matters (Preview)
For due diligence marathons, regulatory deep-dives, and matters that need exhaustive coverage: XL by jhana runs multiple parallel research agents with self-critique and completeness loops.
Meta-agent architecture: multiple Paralegal instances work in parallel
Automatic sub-task decomposition: complex queries broken into focused threads
Scratchpad with synthesis: findings aggregated and deduplicated
Self-introspection: AI critiques its own work, identifies gaps, rebuts assumptions
Citation preservation: every claim traced to source across the entire run
Built for overnight runs: let it work while you sleep
Early Preview: XL is in active development. Available for select users—contact us if you need exhaustive research capabilities for complex matters.
Suit: Citation, Lexicon & Navigation Improvements
Footnotes and citations now work reliably across all document types
Lexicon chip double-click to deselect and remove PDF reference
Automatic viewer switching when selecting lexicon items
Multi-select in Lexicon with next/previous navigation buttons
First five items visible even when categories are collapsed
Scrollbars appear only on hover—cleaner interface
History sources now properly stream via WebSocket
Razorpay Billing: Seamless Indian Payments
Recurring subscriptions and one-time payments now work smoothly for Indian users. Better success rates, better dashboards, and discount campaigns.
Recurring subscription management with auto-renewal
Higher payment success rates for Indian cards and UPI
Campaign discounts and promotional pricing support
Improved billing dashboard with clear status
November 2025
Vision AI: Read Page Ranges as Graphs
For tables, charts, handwritten notes, and visually complex documents that OCR can't handle—Vision AI reconstructs them as structured graphs.
New Paralegal tool: Read Page Ranges' Graph
Full visual understanding of tables, diagrams, and handwritten content
Supplements read() tool for areas where text extraction fails
Powered by Gemini 3 Pro for maximum visual accuracy
Denser, higher-information graphs with increased node detail
Use case: Annexures with complex tables, handwritten Panchnamas, income tax computation sheets—anything where the visual layout carries meaning.
Gemini 3 Pro & Claude Opus 4.5 Now Available
The model picker now includes the latest frontier models. Choose the right model for the task—faster reasoning, bigger context, or specialized legal understanding.
Gemini 3 Pro: 2M token context, exceptional visual understanding
Claude Opus 4.5: nuanced reasoning and careful document analysis
Model picker in Paralegal: switch models mid-conversation
Automatic context handling for long documents across all models
Blog & Case Studies Launch
Real practitioner stories, technical deep-dives, and strategic perspectives on legal AI. With full SEO, social sharing, and CMS infrastructure.
Case studies from lawyers, CAs, tax professionals, compliance officers
Custom article layouts with citation support and source linking
Social sharing with proper OG images and metadata
Full-text search across all articles
RSS and sitemap for discovery
Suit: Source Preview & Text Search Improvements
Source on-click now opens to correct page with highlight
Left and right panel previews with metadata render
Improved text search highlighting and result density
Lexicon collapse by default when ≤1 connection per tag
Color change for visited Lexicon entries
Option to permanently remove certain Lexicon entries
Paralegal Stability & Mobile Improvements
Fixed phantom continuation when context was proactively collapsed
Improved mobile responsiveness with proper 3-panel config
Mobile source views now work correctly
Fixed overflow and padding issues on phone screens
Open in Paralegal button now works from Suit header
Improved prompts for multi-tool passes and multiple sources per claim
Fetch Link Tool Improvements
Better handling of external URLs, improved content extraction, and more reliable source attribution for web content.
Earlier Updates
For updates before November 2025, including the original Suit launch, Paralegal GA, and foundational features, check our archive or contact us. We've been shipping continuously since launch—the changelog above covers our most recent and impactful releases.
KEY TAKEAWAY
We ship weekly. Have a feature request or workflow you wish jhana supported? We read every message—reach out via the in-app chat or email us at hello@jhana.ai.
Contents
Topics
Continue Reading
Discussion
Comments • Share your thoughts and questions below