Update, November 2026: This post describes the previous version of AI Magicx. Its plans, prices, document limits and Document API no longer apply. Today you can attach PDFs, Word and text files to any chat or add them to a project's knowledge, and answers cite the file and page. See pricing for today's plans and the API docs for today's API.
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 AI Magicx, 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
The Document API endpoints announced here were part of the previous version and have been retired. Today's API (OpenAI-compatible chat, runs, media jobs and webhooks, paid from a prepaid API balance) is described in the API docs.
Additional Improvements
Better Usage Tracking
We improved the Usage page to show more detail about document operations:
- Processing costs: See what document processing cost
- Chat costs: Track document question costs separately from regular chat
- Document name: Know exactly which document each charge relates to
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 launched on the paid plans of the time. Those plans have since been replaced; see pricing for today's plans.
Try It Now
Ready to chat with your documents?
- Open a chat, or a project if you want the files kept as its knowledge
- Attach a PDF, Word or text file (up to 25 MB)
- Start asking questions
That's it. Your documents become conversations.
Questions or feedback? We'd love to hear from you:
- Email: support@aimagicx.com
- Twitter: @aimagicxai
Thanks for building with AI Magicx. Here's to a productive 2026!
— The AI Magicx Team