Changelog

biggest update yet, let me cook 🧑‍🍳

• this update is finally live on mobile and the website. be sure to upgrade 📀
• going to be a wall of text, but please read through to understand whats new 🥺
• featuring up to 30+ new features, improvement of existing features, cleaner and more aesthetic ui for both web and mobile, bug fixes and much more
• super excited for folks to try it out ❤️

introducing agent mode, fast mode and library mode

• agent mode allows the model to automatically pick and choose the tools it needs to perform a tasks.
• no longer do you need to manually pick the tool needed for specific capabilities, the agent will automatically pick the tool needed for the task at hand
• also as part of this agent mode, a fast mode is available at your disposal. in fast mode, you are in control of which tools are selected
• library mode is available within library chat where all you want to do is chat with the sources without the distraction of other tools

new computer tool

• added new computer tool to the agent. if you ever wanted the agent to run code for validation, graph visualisation, create docx, pptx, and more, computer tool is right up your alley
• now each agent in your chat has access to a running computer attached to it to perform tasks
• this is automatically enabled once agent mode is selected and the agent will leverage it as required.

video generation improvement

• added Veo 3.1 Fast, Veo 3.1 WAN and xAI Grok Imagine Video generation models
• different video models has different video tasks supported including text to video, image to video, reference to video, video extensions and video editing feature.
• video generation feature is now available for both plus and pro users. some users will require on demand credits for video generation
• video generation cost significantly more credits, be sure to use it wisely

introduce motion tool for agent

• added motion tool to the agent. agent can now create animations for you. this is only available for pro users

web

• major ui and qol upgrade to the web site
• added keyboard shortcut to create new chat
• added AI response transcription to live mode
• added slash / command in chat so you can easily toggle tools, conversation mode, change model and more without leaving your keyboard

mobile

• major ui and qol upgrade to the mobile app
• added ai response transcription to live mode
• live mode continues to work even if the app is in background mode. similar to a phone call
• more tools available to the mobile app

introduce labels

• labels are now available for chats, project library, note, workflow, and whiteboard. You can now better organize your content through labels
• with the introduction of labels, folders will be deprecated soon, please migrate over to use labels

revamp of library feature

• its now called project library, why 🤔 ? because unlike the previous library where it only works when there are sources uploaded, you can now create chat within project library without any sources
• added instruction for project library, this means that you can now set default instruction that the model will follow
• introduce library mode in chat. unlike the old library feature where you are not able to use any tools, you can now pick between fast mode, library mode or agent mode which allow agents to use tools as well within library chat. library mode is selected to default where the agent only has project library and knowledge tool to answer questions based on the uploaded sources
• you can now toggle what sources are included in the library chat as well

introduce project library artifact

• added library artifact feature which allow you to create artifact based on the selected sources
• the current supported artifact are, podcast, quiz generation, flashcard generation, phone call with sources, infographic, mindmap, reports.
• i love using infographic with multiple sources, this is a game changer to how i understand materials
• all this artifacts can be shared publically with the world

introduce workflow

• added workflow, an infinite canvas for your imagination
• llm model invocation, creative tools, from text to image to video generation to motion, you can now branch your logic as you see fit
• with workflow, you can create infinite nodes to connect your ideas and the models output as you see fit. chain em nodes!

on demand credits

• you can now purchase more on demand credits (which are credits that does not expire)

set up default model easily

• one of the most requested feature, a way to set default model quickly
• setting default model is available for both plus and pro users on mobile and web
• on the web, you can now tap on the like icon to set a default model
• on mobile, you can long press the model and choose default model

improvement to notes and whiteboard

• more consistent syncing mechanism
• support for image upload for notes on the mobile app, hooray!

improved live mode

• live mode has lower latency now, better audio interruption and supports background call on mobile

revamp the creative feature ui

• moved the creative tooling on the web to the right panel making it easier for you to visualise
• the mobile creative tooling ui is also updated to reflect the new design language

sharing is caring

• chat can now be shared publically
• project library artifacts created can also be shared publically
• added a public link in the settings where you can manage what is shared

introducing prompt managment tool inside chat

• added prompt management tool inside chat. you can now save prompts and reuse them later within the same chat interface
• tapping on the saved prompts automatically insert that promp into your text area input, how revolutionary

improved web search

• as always, imminment web search improvement for speed and quality of the search

better deep research tool

• as always, imminment deep research improvement for speed and quality of the search, able to reason and orchestrate through more sources to provide more concise, relevant information

revamped memory tool

• introduce a new memory tool, memory are now injected for each conversation thread so that the model has full context about you
• you can choose to toggle this off for each conversation thread
• update and delete memory feature is now available in the setting page. you are in full control of the memory

improved focus os

• improved focus os tab selection on both web and mobile

added pinned chat

• you can now pin chat on the sidebar 😱

updated credit system

• model cost are now based on input and output tokens instead of flat rate like previous versions. this makes credit cost much more accurate based on the task the model actually used and we find this saved the credit usage by 25% for most users.
• this means that credit consumption are more accurate based on your usage and can last longer
• also introduce new credit for each user plan, now all plans have 100x the previous existing credit for respective plans.
• most credit cost from the model to the creative tools are also updated to reflect the new credit system

better tools and coming to mobile

• ai quiz generator, ai flashcard generator, podcast generator, image to prompt are improved for better output
• tools have arrived on mobile, featuring ai quiz generator, ai flashcard generator, podcast generator, image to prompt, ai rewriter, sentence shorter and more

faster upload and download speed

• some of you might notice faster upload and download speed for your images and document compared to previous version, and some of you might not, thats how the orbital world works (something something, cdn 🌎)

new models, hooray

• added xAI Grok 4.1 Fast to the supported model list
• added Google Gemini 3.1 Flash Lite to the supported model list
• added Inception Mercury 2 to the supported model list, fastest Diffusion Based LLM Model
• new upscaler image model
• new replace object image model
• added Nano Banana 2 to the supported model list
• added Black Forest Labs Flux 2 Max
• added OpenAI GPT Image 1.5
• added OpenAI GPT 5.4

free lifetime access to focusbff app for paying zemith user

• in case you miss it previously 🔐

enjoy!!!

• there might be stuff i've missed in this changelog, if you see something new lmk
• this is month of hardwork in the background, bringing major revamp and improvement to the tool that so many are using on a daily basis, hope you enjoy it