Changelog
This document lists all recent MythWeaver updates, the changes contained therein, and other developer notes.
v2.20
Image Generation Changes
Kontext Editing: Premium feature that lets you modify existing images with contextual awareness - tell it what you want changed and it intelligently edits the image while maintaining consistency with the rest of the scene.
Rocky Integration: Added a new backend for image generation that handles requests more efficiently, improving reliability and speed for all image creation.
Image-Based PDF Extraction: You can now upload PDFs with images and we'll extract the visual content along with text, making it easier to pull inspiration and information from your existing campaign materials.
Image Feedback System: Added the ability to provide feedback on generated images to help us improve quality over time.
Weavy Improvements
Multi-Tool Support: Weavy can now use multiple tools simultaneously in a single response, making conversations significantly faster and more efficient when generating complex content.
Multi-Lore Creation: Create multiple lore entries from a single Weavy conversation, perfect for quickly populating your campaign world.
Big File Support: Improved handling of large file uploads in Weavy conversations, so you can drop in bigger PDFs and documents without issues.
New Features & Improvements
Campaign Switching Without Refresh: Switch between campaigns instantly without the annoying page reload. Because waiting for pages to load is some 2010 bullshit.
Dynamic Mention System: Completely rebuilt the @mention functionality to be more reliable and intuitive when referencing your lore in conversations.
In-App Password Changes: Email/password users can now change their passwords directly in the app settings instead of going through the password reset flow.
Arrow Key Navigation: Use arrow keys to navigate through search results like a goddamn power user.
Relationship Deletion Confirmation: Added a confirmation dialog when deleting relationships between lore entries to prevent accidental deletions.
Dynamic Template Costs: Template generation now uses variable credit costs based on complexity and expanded token outputs, giving you more detailed content where it matters.
Conversation Title Updates: Can now edit conversation titles in Weavy to keep your chat history organized.
New Establishment Template: Added a dedicated template for taverns, shops, guilds, and other establishments with fields that actually make sense for these locations.
3D Model Image Reset: Added ability to reset the uploaded reference image for 3D model generation if you want to try a different reference.
Behind-the-Scenes
New Favicon: Because we were tired of looking at the old one.
SD 3.5 Full Release: Stable Diffusion 3.5 is now fully available to everyone (removed the feature flag).
v2.19
Template Generation Overhaul
Async Template Generation: Template creation is now significantly faster because it runs in the background. Start generating a character or location and keep working while it processes.
All-at-Once Generation: Instead of generating templates field-by-field, we now generate the entire template in one shot, which is faster and produces more coherent results.
Drastically Improved Content Quality: Lore generation now produces richer, more detailed content with proper formatting, including stat blocks for characters and creatures that follow standard TTRPG formatting conventions.
v2.18
Image Generation Changes
V2 Image Model Support: Complete overhaul of how we handle image models, adding support for next-generation versions of existing models with better quality and features.
Auto-Select New Model Versions: When a model gets updated, the system automatically selects the newest version so you always get the best quality without having to manually switch.
Improved Model Selection: Better handling of model families and version descriptors to make it clearer which model you're actually using.
Mobile Improvements
Optimized 3D Model Viewing: 3D model viewer now works much better on mobile devices with improved performance and touch controls.
Mobile Image Generation Cleanup: Streamlined the image generation interface on mobile to be cleaner and more intuitive.
Simplified Lore Creation: Creating empty lore documents on mobile is now simpler and more straightforward.
New Features & Improvements
Collection Move Improvements: Completely cleaned up the functionality for moving entire collections and conjurations between folders, making organization way more reliable.
Max Credits for Free Plan: Free users now have a credit cap to keep system usage sustainable.
Fixed Conjuration Visibility: Resolved issues where some conjurations weren't showing up in collections.
v2.17
Image Generation Changes
Premium Inpainting: High-quality inpainting is now available for premium subscribers, letting you edit specific parts of images with better quality and more control than the standard tools.
New Features & Improvements
UI Standardization: Rolled out consistent empty states and standardized UI patterns across the app so everything feels more cohesive.
Improved Feature Gating: Better handling of feature gating for premium features with clearer messaging about what's available at different subscription tiers.
v2.16
Weavy Improvements
Create Lore from Conversations: You can now instantly convert Weavy conversations into lore entries, making it easy to capture worldbuilding ideas that emerge during planning sessions.
Campaign-Specific Weavy: Weavy conversations are now tied to specific campaigns, so your chat history stays organized by campaign.
New Features & Improvements
Guest Lore Creation: Campaign guests can now create lore entries, making collaborative worldbuilding actually collaborative.
Better Error Recovery: Improved how the app handles and recovers from errors to prevent crashes and data loss.
v2.15
Image Generation Changes
Flux 1.1 Pro Model: Added Black Forest Labs' Flux 1.1 Pro model, delivering exceptional image quality with better prompt following and more detailed outputs.
v2.14
Image Generation Changes
Stable Diffusion 3.5 Flash & Large: Added two new Stable Diffusion 3.5 models - Flash for faster generation and Large for higher quality - both with significantly better prompt understanding than previous versions.
Model Selection Overhaul: Completely rebuilt the model selection interface to be more compact and easier to navigate, especially with the growing number of available models.
Billing & Subscription Changes
Credits System Overhaul: Completely rebuilt how credits and complimentary access work to be more transparent and reliable.
Next Credit Allocation Date: You can now see exactly when your next credits will be allocated so you're never wondering when your monthly credits refresh.
Payment Method Display: Shows your saved payment methods when upgrading, making subscription changes clearer and easier.
v2.13
Image Generation Changes
Improved Error Handling: Made image generation error messages much clearer so you actually understand what went wrong instead of getting cryptic error codes.
Prompt Length Clarity: Better communication about prompt length limits and requirements to help you craft better prompts on the first try.
v2.12
Image generation changes
We made some foundational improvements which let us now support up to 2000 characters for image model prompts, double what we previously supported.
We now handle image generation errors more consistently across all types of errors and the app should communicate that to users more clearly now.
Performance & Reliability Improvements
Infrastructure Enhancements
Enhanced Security & Monitoring: Upgraded infrastructure to protect user data and made general fixes to improve security posture.
Speed & Performance
Made a number of changes to help make working with MythWeaver a smoother, faster and overall more enjoyable experience. Page load times across the board should now load at least 2x faster and even faster than that in some cases.
Behind-the-Scenes Fixes
Smoother File Handling: Improved how the system processes and stores files for better reliability.
More Consistent Interface: Fixed various small visual inconsistencies to provide a more polished experience throughout the app.
v2.11
New Features & Improvements
Enhanced Weavy Conversations
Upload Documents to Weavy: You can now upload documents (PDFs, Word docs, text files) directly to your Weavy conversations. Weavy will read and understand your documents, giving you more accurate and contextual responses based on your uploaded content.
Campaign File Integration: When chatting about your campaigns, the AI can now access and reference files you've uploaded to that campaign, providing more relevant suggestions and responses that understand your world's context.
Full Screen Weavy: You can now toggle Weavy into full screen mode to make better use of all the pixels available to you.
Beautiful Message Formatting: Weavy responses now display with proper formatting including headers, bullet points, code blocks, and other rich text elements, making responses easier to read and more professional-looking.
Copy Responses Instantly: Added a convenient copy button to AI messages so you can easily grab responses and use them in your documents or share them with others.
Performance & Reliability Improvements
Behind-the-Scenes Fixes
Smoother File Handling: Improved how the system processes and stores your uploaded files for better reliability.
More Consistent Interface: Fixed various small visual inconsistencies to provide a more polished experience throughout the app.
Better Error Prevention: Enhanced systems to prevent conflicts between different parts of the interface.
Last updated