Changelog

The Roleplay mini-app has been completely redesigned: https://www.layla-network.ai/post/roleplay-mini-app-redesign

Layla supports local voice cloning with PocketTTS! https://www.layla-network.ai/post/layla-v6-11-0-has-been-published

Create your own voices by recording or uploading an audio sample. Layla will create a custom voice out of it for which you can use for your characters! You can also set your custom voice as the the global TTS voice for Layla, or configure it individually for each character.

Record audio right in Layla to create custom voices: you can record an audio snippet and use it as a base to clone a voice.

Voice Cloning tips:

PocketTTS is very sensitive to background noise. Record your audio in a quiet room or use an online editor to remove background noise to get better results.

Only supports 16-bit PCM wav files, use an online converter if you need to convert mp3's etc.

Try not to have blanks/silences in the sample, PocketTTS will try to learn that and generate lots of silences.

Exaggerated voices work better.

Layla x Local-Dream! https://www.layla-network.ai/post/layla-v6-10-0-has-been-published

The creator of Local-Dream (https://github.com/xororz/local-dream) has graciously allowed direct integration with Layla.

Introducing "Companion Mode" in Layla. (Android only.)

Read more here: https://www.layla-network.ai/post/layla-v6-9-0-has-been-published

This update bring Gemma-4 support, improves agent functionality, with a new Choose Your Own Adventure agent, and improves chatting UI with more customisations. Read more here: https://www.layla-network.ai/post/layla-v6-8-0-has-been-published

This update fleshes out Layla's Agent Framework.

This update adds support for LiteRT-LM models, improves model loading times, and contains numerous improvements and bug fixes. Read more here: https://www.layla-network.ai/post/layla-v6-6-0-has-been-published

This update redesigns the character selection screen, giving you detailed information on their statistics such as number of messages and past memories. Read more here: https://www.layla-network.ai/post/layla-v6-5-0-has-been-published

This update redesigns the long-term memory graph. It is now in 3D! If you'd like to learn more about the visualisation of your memories, read here: https://www.layla-network.ai/post/layla-v6-4-0-has-been-published

Layla can now replace Google Gemini as your default assistant! You can do numerous tasks on your phone such as drafting emails, creating reminders etc. without opening the main app by simply holding the power button.

Configure this by going to Android System Settings -> Digital Assistant -> choose Layla.

Layla 6 supports Agents!

Read more about the updates here: https://www.layla-network.ai/post/layla-v5-6-0-has-been-published

Read more about the updates here: https://www.layla-network.ai/post/layla-5-5-0-has-been-published

Read more about the updates here: https://www.layla-network.ai/post/layla-v5-4-0-has-been-published

Details on the updates are here: https://www.layla-network.ai/post/layla-v5-3-0-has-been-published

Read in detail the new features here: https://www.layla-network.ai/post/layla-v5-2-0-has-been-published

Read in detail the new features here: https://www.layla-network.ai/post/layla-v5-1-0-has-been-published

Read in detail the new features here: https://www.layla-network.ai/post/layla-v5-0-0-has-been-published

This update adds SillyTavern Lorebook support! https://www.layla-network.ai/post/layla-v4-16-0-has-been-published

This update brings support for the i8mm CPU feature. This makes LLMs run much much faster on latest CPUs! Note this requires a special quant called "048". You can read more about how to get them and if your phone supports them here: https://www.layla-network.ai/post/layla-supports-i8mm-hardware-for-running-llm-models

Details here: https://www.layla-network.ai/post/layla-v4-14-0-has-been-published

Details here: https://www.layla-network.ai/post/layla-v4-13-0-has-been-published

This update contains a significant UX improvement when switching models: https://www.layla-network.ai/post/layla-v4-12-0-has-been-published

The highlight for this update is Layla's Toolkit upgrade to v2: https://www.layla-network.ai/post/layla-v4-11-0-has-been-published

This version contains the Lorebook feature, where you can attach multiple documents to your characters to act as background information during your chats: https://www.layla-network.ai/post/layla-v4-10-0-has-been-published

This version contains numerous new features including a completely redesigned voice chat!

Read the detailed updates here: https://www.layla-network.ai/post/layla-v4-8-0-has-been-published

Read more about all the different features here: https://www.layla-network.ai/post/layla-v4-7-0-has-been-published

Read more about all the updates here: https://www.layla-network.ai/post/layla-v4-6-0-has-been-published

Read more here: https://www.layla-network.ai/post/layla-v4-4-0-has-been-published

Full release notes are here: https://www.layla-network.ai/post/layla-v4-2-0-has-been-released

This update brings various quality-of-life improvements that are requested by the community.

In particular, you can now add/edit/delete long-term memories for individual characters! A new "small grid" display option is added for the character selection screen, with the ability to drag & drop to re-order characters.

This update adds Claude API support! With Claude 3's release, the general consensus is that it is better than ChatGPT overall! (See the Chatbot Arena Human Eval).

Additionally, this update focuses on continuing to provide stability updates and resolves some of the issues with image recognition reported here.

This is a stability update that addresses numerous issues reported after the v4 update.

Read more about the update here: https://www.layla-network.ai/post/layla-v4-0-0-has-been-published

In this update, we added a hilarious new app: Anarchy Chess!

We all know LLMs are terrible at chess, you can now experience this first hand by playing with all your favourite characters.

But the main strength of the LLM lies in the fact that it's a language model. Which means it will trash-talk and gaslight you throughout the game 💀.

Demo: https://youtu.be/tfoYJ3nzPgw

Read more on our blog here: https://www.layla-network.ai/post/layla-v3-12-0-has-been-published

Read the details on our blog: https://www.layla-network.ai/post/layla-v3-11-0-has-been-published

Demo here: https://youtu.be/FWpl_W7TOhA

Read the details on our blog: https://www.layla-network.ai/post/layla-v3-10-0-has-been-published

Demo here: https://youtu.be/nijMsEsAjfs

Read the details on our blog: https://www.layla-network.ai/post/layla-v3-9-0-has-been-published

Read the details on our blog: https://www.layla-network.ai/post/layla-v3-8-0-has-been-published

Read in detail about this patch on our blog: https://wix.to/kmjRmij

Demo video of KIP with text to speech (sound on): https://youtube.com/shorts/1S0KiqflBtA?feature=share

Read in detail about all the updates in this patch and how they tie into the overall roadmap on our blog: https://www.layla-network.ai/post/layla-v3-6-0-has-been-published

Demo video of Annie (new character): https://youtube.com/shorts/qci-66GJeC4?si=XGpv9PfjgWQNTY8T

Highlights for this update are improvements to the RMI app (Raw Model Instructions)!

You are now able to adjust prompt templates used for your custom models. ChatML and Vicuna are given as examples, but the configuration should really support any format.

Also in the RMI app, you can manage storage and delete the custom models copied to your app's cache. This should help those of you who like to rapidly switch between lots of different models.

In answer to all your feedback regarding proactive messaging, "Proactive Chat" has been moved to its own app. This gives you fine-grained control over how you want it to work. You can select the frequency as usual, in addition, you can also choose which characters are allowed to proactively chat with you. Removing the "Proactive Chat" app stops all future notifications from being sent to you; adding it back re-enables the notifications. I hope this addresses the majority of use-cases!

Note: In this update proactive chat has been turned off for everyone. If you would like to re-enable it, please add the app to Layla in the App Screen.

Lastly, this version continues to lay the groundwork for long term memory, including a locally implemented vector database and embedding model with sub-1-second performance.

As the number of features ("apps") in Layla increases, a new way to manage them is introduced! "Features" in Layla now works like a "mini app store": you can select which apps you want to include and remove them if you don't need them.

Adding an app enables the corresponding feature, and removing it disables it. For example, if you add the "To-do" app, Layla will pick up scheduling instructions, and once you remove the app, Layla will ignore your scheduling instructions. The "mini app store" replaces the individual "enable/disable" toggle of apps.

Note: with this update, all your apps are disabled by default, you can add them back using the browse app screen. There's no additional download for each app.

Changed the default sampling method for Layla to Dynamic Temperature Sampling. See the paper: https://arxiv.org/abs/2309.02772

This sampling method is reported to provide more organic responses from models by various users in the local AI community. It has already been implemented in other popular frontends such as Oobabooga, KoboldCpp, etc., and is scheduled to be added to SillyTavern as well, so it as seen a fair number of real-world use-cases.

You can turn this feature off by setting "Dynamic Temperature Range" setting in the Advanced Settings page to 0.

Added a new app: "My Horoscopes".

This app gives you the latest horoscope information for each day. You can then chat with Layla about your horoscopes. Layla will give you tailored advice taking into consideration your horoscope for the day. You can also optionally enable notifications each morning to send you your horoscope information. As always, please don't take advice generated by an AI seriously.

Your personal information (such as your zodiac sign or birthday) is not leaked. This app works by downloading information for ALL horoscopes first, then you can choose which one is yours locally on your device.

Horoscopes may seem a frivolous little feature at first glance, but this is a proof-of-concept for two important operations:

Real-time information retrieval using the internet.

Augmented generation using relevant info (also known as RAG)

This is laying the foundation for a very important update later: the ability for Layla to search and verify information on the internet.

This update contains various quality-of-life improvements.

Layla now supports any GGUF models on the internet!

You can download models from your own sources and load them into Layla. A new app "Raw Model Instructions" is added where you can configure the settings for your own GGUF models. While the inbuilt "Layla" models are still recommended for general use, any sufficiently intelligent LLM model should be able to pick-up the prompt style of Layla relatively quickly after a few conversations and may achieve better results in specific areas such as storytelling, roleplay, etc. This is an advanced feature; use at your own discretion. Loading a model too large for your phone memory will crash without warning.

This update comes with completely new base models for both the Lite, Q2, and Full versions! After updating, the app will start a new 2-4GB download! (Legacy models will not be deleted. You can switch back to them in the settings page if you wish.)