What's New in NovaKit: January 2026
Introducing Document Chat—upload PDFs, URLs, and YouTube videos, then have AI conversations with your content. Plus collections, citations, and more.
What's New in NovaKit: January 2026
Happy New Year! We're kicking off 2026 with a major new feature: Document Chat. Upload your PDFs, paste URLs, import YouTube videos—then have intelligent conversations with your content.
Let's dive into everything new this month.
Document Chat
The headline feature of this release. Document Chat brings RAG (Retrieval-Augmented Generation) to NovaKit, letting you get AI-powered answers grounded in your own documents.
Four Ways to Add Content
1. File Upload Upload PDFs, Word docs, Markdown, text files, and CSVs directly. Drag and drop or click to browse.
2. URL Import Paste any public URL and we'll extract the content automatically. Great for articles, documentation, and blog posts.
3. YouTube Import Paste a YouTube video URL to import its transcript. Chat with lecture content, tutorials, or podcasts without watching hours of video.
4. Paste Text Paste text content directly—meeting notes, emails, code snippets, or anything else.
Intelligent Answers with Citations
Every AI response includes numbered citations linking back to the exact source passage. Click any citation to see the original context.
Revenue grew 35% year-over-year [1], primarily driven
by enterprise expansion [2].
No more "trust me" answers. Every claim is verifiable.
Multi-Document Chat
Chat across multiple documents simultaneously. Upload related files, select them all, and ask questions that span your entire knowledge base.
Example: Upload a contract and all its amendments, then ask "What are the complete payment terms across all documents?"
Collections
Organize documents into collections for:
- Project-based grouping
- Topic categorization
- Team collaboration
Chat with an entire collection at once. Perfect for research, competitive analysis, or managing large documentation sets.
Processing Progress
Watch your documents process in real-time:
- Parsing → Chunking → Embedding → Ready
- Progress bars with ETA estimates
- Stage-by-stage status updates
Retry Failed Imports
URL or YouTube import failed? Click Retry to try again without re-uploading. We improved error messages so you know exactly what went wrong.
Document API
Full API access for Document Chat is now available on Pro and Business plans.
New Endpoints
Upload documents
POST /api/v1/documents/upload
Import from URL/YouTube
POST /api/v1/documents/import
{
"source_type": "url",
"url": "https://example.com/article"
}
Chat with documents
POST /api/v1/documents/chat
{
"document_ids": ["doc_123"],
"message": "What are the key findings?"
}
Streaming responses Server-sent events for real-time chat responses, just like our AI Chat API.
Check the full documentation for examples and SDKs.
Additional Improvements
Better Credit Tracking
We improved the Usage page to show more detail about document operations:
- Embedding costs: See how many credits document processing used
- Chat costs: Track RAG query costs separately from regular chat
- Document name: Know exactly which document each charge relates to
Document Limits
The Usage page now shows your document quota alongside other limits. Pro users get unlimited documents; Free users can try with up to 5.
Improved YouTube Transcript Extraction
We upgraded our YouTube transcript extraction to be more reliable:
- Better language fallback (tries multiple caption tracks)
- Improved error messages when transcripts aren't available
- Faster extraction for long videos
Pre-Flight Checks
Before creating a document record, we now verify:
- URL is accessible and not behind authentication
- YouTube video has available captions
- File format is supported
This means fewer "failed" documents cluttering your library.
Pricing
Document Chat is available on Pro ($19/mo) and Business ($59/mo) plans.
Existing Pro and Business subscribers get Document Chat automatically—no action needed.
What's Next
We're already working on the next wave of document features:
- OCR support: Process scanned PDFs and images with text
- Document summaries: Auto-generated summaries when you upload
- Shared collections: Collaborate with team members on document sets
- Knowledge agents: AI agents with access to your document collections
Stay tuned for updates.
Try It Now
Ready to chat with your documents?
- Go to AI → Documents
- Click Add Source
- Upload a PDF, paste a URL, or import a YouTube video
- Start asking questions
That's it. Your documents become conversations.
Questions or feedback? We'd love to hear from you:
- Email: support@novakit.ai
- Twitter: @novakitai
Thanks for building with NovaKit. Here's to a productive 2026!
— The NovaKit Team