A lighter update this week, with a handful of fixes across web and mobile.

Recommendations fixed on Android

We fixed an issue where the Recommendations page on Android was showing the same chatbots as the homepage instead of your personalized For You recommendations.

Persona avatar generation restored

Free and Get a Taste users can once again use the Generate option for Persona avatars on Desktop and Mobile Web. Previously, only the Upload option was appearing.

Other fixes

• Fixed chatbot rating counts showing as missing or zero for some characters on the homepage.
• Fixed Anime/Semi-Realistic V2 image generation ignoring a 0 LoRA scale setting and resetting it to 0.8.
• Fixed an Internal Server Error appearing when detaching a Lorebook from a chatbot.
• Fixed an error when uploading animated WebP images as chatbot avatars.

That’s it for this week. Thanks as always for the reports and feedback. 💜

A few useful updates went live this week, including model changes, quality-of-life improvements, and fixes across the platform.

Zeta G4 is now part of the Default model selection

Since its release, Zeta G4 has become a favorite among Free and Get a Taste users and has been performing well across the platform.

We’ve now added it as one of the models used under the Default selection. Upgrading the default experience from 8B to 26B model. 

Since the default model changes now and then, Zeta G4 is also still available separately, so you can select it directly whenever you specifically want to use it.

We’ve also made the previous default model, Lyra Spiced, available as its own option for anyone who preferred its style. It’s a fine-tuned version of Lyra that works well for lighter, general-purpose roleplays.

 Other improvements/features 

• You can now sort chats by Recent, Oldest, or Most Active.
• Restored Sort by Name on the My Chatbots page.
• Voice playback is now hidden on bot messages longer than 2,000 characters, since TTS does not support messages beyond that length.
• One of the most frustrating issues was losing a message you had spent time writing when an error occurred. Failed messages will now remain saved as a draft, so you can try sending them again instead of starting over.

Bug fixes

• Fixed an issue on Android where chatbot cards displayed every assigned tag instead of a limited number.
• Fixed an issue on Android where the Copy Message option could be cut off on longer messages.

Beta tester applications closing soon

We’re also preparing some new features, models, and experiments for our beta testing group.

Interested in getting early access and sharing feedback directly with the team? 

Apply Here

Applications close this Sunday.

As always, thank you for the feedback and reports. 💜

Hey everyone 👋

Nothing too major this week, but we do have one requested improvement for creators.

Small edits now apply instantly

Some creator edits will now apply immediately instead of going through review.

This includes edits limited to:

• Tags
• The Exclude from Group Chats toggle
• The Show Personality toggle
• Attaching or removing a voice
• Attaching or removing a Lorebook

Our priority is to keep creating on SpicyChat fun and smooth, while still keeping the platform safe. This should make smaller updates much easier without adding unnecessary waiting time.

Also, spoiler alert for free users: something nice is cooking for you. 👀

A smaller update this week, but we’ve pushed a few useful fixes and improvements across web and mobile.

Russian payment methods with VPN

Russian users should now be able to access Russian payment methods even when using a VPN.

On the checkout page, set your Billing Country to Russia, and the available Russian payment methods should load correctly.

Web fixes

• Fixed an issue where the Discard Changes confirmation did not appear on the Create and Edit Persona pages after making changes and trying to leave without saving.
• Fixed an issue where the Suggest Tags dropdown on chatbot profiles would stay open after clicking outside of it.
• Fixed an issue where some buttons on the Add Member page on group chat creation were not translated properly when using another site language.
• Fixed an issue where the Auto Suggestion button on the Create Group page was not generating content for the Greeting and Scenario fields.

Mobile app fixes

• Fixed an issue where the Generate Image modal could not be scrolled properly while the keyboard was open, making some fields hard to access while typing.
• Fixed an issue where the Unsaved Changes popup appeared on the Create Chatbot page even when no changes had been made.

That’s it for this week.

As always, thank you for the reports and feedback. 💜

Hey everyone! Here's a roundup of improvements and fixes we shipped this week — a lot of it driven directly by your feedback. 🙌

But before that, I just wanted to share...

Since its release just one week ago, GLAM 5.1 has quickly become a fan favorite. Users are calling it our most immersive, most human-feeling model yet — with characters that stay believable, stories that drive themselves, and writing that genuinely feels alive.

If you haven't tried it yet, you're truly missing out.

👉 𝐆𝐞𝐭 𝐆𝐋𝐀𝐌 𝟓.𝟏  

The weekly update now:

🔍 𝐒𝐞𝐚𝐫𝐜𝐡 & 𝐁𝐫𝐨𝐰𝐬𝐢𝐧𝐠

✅ 𝐒𝐨𝐫𝐭𝐢𝐧𝐠 𝐧𝐨𝐰 𝐬𝐭𝐢𝐜𝐤𝐬. When you pick Latest or Popular, results stay sorted that way. Previously, relevance ranking could quietly override your choice — that's fixed, so browsing should feel a lot more predictable now.

✅ 𝐃𝐞𝐟𝐢𝐧𝐢𝐭𝐢𝐨𝐧 𝐒𝐢𝐳𝐞 𝐟𝐢𝐥𝐭𝐞𝐫𝐬 𝐮𝐧𝐬𝐭𝐮𝐜𝐤. We fixed a frustrating bug where you could end up with multiple Definition Size filters selected and be unable to deselect them after certain page changes. You can now toggle them freely again.

🎨 𝐀𝐯𝐚𝐭𝐚𝐫𝐬

✅ 𝐅𝐞𝐰𝐞𝐫 𝐟𝐚𝐥𝐬𝐞 "𝐫𝐞𝐚𝐥 𝐩𝐞𝐫𝐬𝐨𝐧" 𝐟𝐥𝐚𝐠𝐬. We resolved the issue where avatars were being wrongly flagged as real people. We've improved the system, so these false flags should be far less common going forward. We know it was annoying — thanks for your patience while we worked on it. 💜

✅ 𝐑𝐞𝐟𝐞𝐫𝐞𝐧𝐜𝐞 𝐢𝐦𝐚𝐠𝐞 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧 𝐟𝐢𝐱𝐞𝐝. We resolved a bug where avatar generation could fail when using reference images. Image-based avatar creation should be much more reliable now.

As always, keep the feedback coming — it genuinely shapes what we fix next. 🚀

We've rolled out several improvements aimed at making the creator experience on SpicyChat smoother, more stable, and more predictable. Here's a clear walkthrough of what's new.

Human Review for Edits to Published Chatbots

Until now, edits made to already-published public chatbots were evaluated primarily through automated moderation. This worked well in most cases, but certain edits could occasionally require a closer look after publishing.

Going forward, edits submitted to already-published public chatbots will also be reviewed by our human moderation team, in addition to automated systems. This adds a second layer of care to the process and helps catch nuances that automated review alone may not.

While your edit is being reviewed, the currently approved version of your chatbot remains live and fully accessible. Your users can continue chatting with it without interruption, and you don't have to worry about downtime while waiting for the new version to be approved.

In short: stronger review, zero disruption.

Group Chat Controls

You now have direct control over whether your chatbot can be added to public group chats created by other users.

This option is available under Advanced Settings in the chatbot editor.

Please note: This setting applies only to future group additions. Chatbots that were already added to group chats before you toggle this setting will remain in those chats and won't be affected.

Lorebook Imports

You can now bring your existing lorebooks from other platforms directly into SpicyChat — no need to recreate them from scratch.

How to use it:

1. Go to the Lorebook creation page
2. Click Import
3. Upload your file and start using it right away

Bug Fixes & Improvements

• Fixed an issue where Lorebook entries attached to private bots were not functioning correctly
• Fixed an error that could appear when starting a Group Chat without selecting who sends the first message
• Added clearer limits and error messaging for Custom Voices, so it's easier to understand when a limit has been reached

One More Thing

Something significant also shipped today. We'll be sharing the full details tomorrow — see if you can spot it in the meantime.

Thank you, as always, for the feedback that continues to shape SpicyChat. We're excited to keep building alongside you.

Improved Image Generation

Excited to let you know that we’ve released an updated version of the Qwen Conversation Image prompt generator.

When generating images in conversation, the underlying prompt should now be much more accurate, which should lead to more relevant and better-looking results.

It also now properly respects Prompt Prefix and Negative Keywords from Image Settings — something that wasn’t working reliably before.

Better Recommendations

On the discovery side, we’ve made a few improvements as well:

You’ll now continue to see Recommendations even when tags are applied, and they will respect the tags you’ve selected.

We’ve also increased the number of recommendations from 12 to 24, so you get more content to explore.

Group Chat Improvements

For Group Chat, a couple of UX improvements:

* The Add/Remove Members button is now easier to access and placed above the selected members list

* Added a Discard Changes confirmation when exiting group member selection with unsaved changes

That’s it for this week.

As always, feel free to share feedback — especially on the image generation improvements.

Hey everyone,

Sharing a quick update on some of the recent fixes and improvements, along with a few things we’re still actively working on.

We’ve made a small but important change to moderation — if your chatbot is in a “Fix Required” state, you won’t be able to chat with it until the required changes are made. This helps ensure bots meet the guidelines before being used again.

We’ve also added a new feature based on your feedback:

Disable Memory

You can now disable automatic memory generation in chats. This means the system won’t add new memories on its own, but any manually added or pinned entries will still be used.

This is useful if you want full control over the narrative and what information is carried forward. That said, for regular use, we don’t recommend disabling it, as it may make conversations feel less coherent.

🐞 Fixes

A number of issues have been addressed to improve day-to-day usage:

• Fixed an issue where reporting a chatbot would sometimes show a translation key instead of a proper success message.
• Fixed an issue where editing and saving a chatbot could still trigger an “unsaved changes” popup.
• Fixed an issue where long memory entries would overflow outside the UI container.
• Fixed an issue where the edit text area wouldn’t scroll properly for longer messages on web.
• Fixed an issue where the favorite button would keep loading infinitely for users who weren’t logged in.
• Fixed an issue where starting a new chat could sometimes carry over messages from a previous conversation.

⚠️ Ongoing Work

We’re aware some users are facing issues with avatar uploads, especially false “real person” moderation flags. We’re actively improving the models to reduce these false positives.

We’re also working on improving GIF uploads — currently files under 5MB should work, and we’re aiming to restore the 10MB limit soon.

That’s it for now — more improvements coming soon.

As always, really appreciate the feedback you’ve been sharing. It helps us move faster and fix things properly. 💜

We’ve rolled out a couple of exciting improvements on the web that give you more control, better flow, and a smoother creative experience.

🎨 Improved Avatar Generation 

We’ve revamped the Avatar Generator with new models and tools to help you fine-tune your characters more easily.

What’s new:

• New powerful image models: Z-Image, Flux2 Klein, and Qwen Image
  
• AI-assisted image editing – make refinements after generating an image to get exactly what you want (like fixing deformities or removing unwanted elements)
• Image cropping to better fit avatars on chatbot cards
• Style picker for easier style selection from a list of predefined styles
• Larger editing playground on desktop
• Improved full-screen image viewer

A couple of important notes:

• These new models work best with detailed, prose-style descriptions. If you don’t describe a background or setting, the image may generate without one.
• NSFW support is limited for AI-assisted image editing. Editing requests involving NSFW body changes may not work as expected.

💬 Cleaner, More Focused Chat Input

We’ve also reworked the chat input to make conversations feel more fluid and less cluttered.

What’s changed:

• The text input is now focused purely on typing and all secondary actions (voice, image upload, suggest reply, commands, web search) have been moved into a new “+” action menu
• Opening or closing the menu won’t affect your typed text
• Actions only trigger when you explicitly tap them — no more accidental interruptions
• Images can now be dragged and dropped directly into chat
• Image-only messages are supported
• You can remove an attached image with a confirmation prompt
• When using commands (like OOC or Search), the input clearly switches into a command mode, making it visually distinct from normal messages

The goal here is simple: less friction, fewer mistakes, and a smoother writing flow.

📌 Availability
These updates are currently web-only and not yet available in native apps.

As always, thank you for the feedback that helped shape these improvements. We’re excited to see what you create with them  💜

Let’s start 2026 with a bang. We’ve got a few exciting updates and improvements to share with you.

📘 Private Lorebooks Now on Android

Private Lorebooks are now fully available on the Android app, bringing the same experience you’ve had on web straight to mobile app.

Make sure you’re on Android version 1.0.88 to access this update.

Public Lorebooks are on the roadmap for 2026, so stay tuned for that. 

⚙️ New Feature: Generation Presets (Web Only for Now)

As we continue to introduce more models that work better with different generation settings, we know switching generation settings manually over and over can get tedious.

That’s why we’re introducing Generation Presets. You can now save your preferred generation settings and switch between them instantly while chatting. 

This feature is for subscribers only.

📖 You can read more about Generation Presets here

🔄 Refresh Recommendations

We’ve added a Refresh Recommendations option for users who feel their recommendations page hasn’t changed in a while.

This lets you:

- Pull in the latest recommendations

- Refresh the list if it feels stuck

Please note: this won’t change anything if you’re already seeing the most up-to-date recommendations.

🛠️ Fixes & Improvements

We’ve also shipped several quality-of-life fixes across web and Android:

- Fixed an issue where the Home page wouldn’t auto-scroll to the top when changing pages.

- Resolved a bug where the Android UI could block the “See More” button on chatbot profiles.

Additional UI and UX improvements to Lorebooks for smoother creation and editing.

That’s it for this update. As always, thank you for your feedback and support — it directly shapes what we build next.

Happy Chatting! 💜