Changelog

New updates and improvements to Zemith.

Today

v1.2.0 major update

biggest update yet 🚀

  • 1.2 is out on web and mobile today. please update the app to 1.2 📲
  • zemith now does the work, not just answers: a smarter agent with its own computer, deep research you can steer, cloud tasks that run while you're away, and plugins for the apps you already use

one smarter agent everywhere

  • no more picking agent mode, fast mode or library mode. there is now one agent, and it decides which tools it needs
  • every chat gets its own computer. the agent can write and run code, install packages, build things and keep working on the same files across messages
  • for bigger jobs you see a live task checklist with progress as it works through each step
  • pick tools, plugins and skills yourself with / in the composer, and tag an image with @ so the agent knows exactly which one you mean
  • save your custom system instructions with "use for new chats" and they follow you across devices

deep research, rebuilt 🔬

  • you now see a research plan first. edit it, start it or cancel it
  • it keeps running even if you close the chat
  • the report comes back with citations checked against the sources

introducing zemith cloud ☁️

  • hand off longer tasks to cloud, close the tab, and come back when they're done
  • tasks keep running in the background for as long as they need. if one needs your input or approval, it pauses and lets you know
  • cloud projects keep files, standing instructions and a project memory that remembers things across tasks
  • schedule tasks to run once, daily or weekly, or hit run now anytime
  • get notified on web and mobile when a task finishes, fails or needs you

realtime across web and mobile ⚡

  • chats now sync live between web and mobile. start a message on your laptop and watch it stream on your phone, or the other way around
  • responses no longer get lost when you refresh or switch devices. the answer keeps generating in the background and picks up wherever you open the chat

files in chat 📁

  • every chat now has a files view showing what you uploaded and what the agent created, side by side
  • preview or download any of them from there
  • ask for a word doc, presentation, spreadsheet or pdf and the agent builds the actual file, with proper layouts, headers, tables and colour palettes

edit images right in chat 🎨

  • hit edit on any generated image, describe the change, send. works on web and mobile

fork and share chats

  • fork a whole chat, or fork from any single response with more → fork from here
  • share a chat with your teammates to work on it together, and see who wrote each message

projects 📂

  • the agent can now open and work with your project files directly, like calculating over a spreadsheet instead of only searching its text

plugins 🔌

  • browse and connect plugins from the new plugin marketplace
  • connect your apps and let the agent work with them: Google Calendar, Outlook Email, Outlook Calendar, OneDrive, Microsoft Word, Excel, PowerPoint, SharePoint, Notion, Linear, Airtable and Canva
  • make short videos from a prompt with the HyperFrames plugin
  • whenever the agent wants to change something (send, create, edit), it asks you first. approve or deny from web or your phone

skills

  • create your own reusable skills and use them in any chat
  • starter skills made by zemith: meeting notes, weekly update, writing polish, decision brief, study guide and skill builder

introducing workspaces 👥

  • everything you do in zemith now lives in a workspace: your chats, projects, files, cloud tasks, plugins and credits
  • you start with your own personal workspace, and can create team workspaces for work with others
  • switch between workspaces anytime with the workspace switcher, on web and mobile
  • invite people into a team workspace. each member gets their own seat and credits, and top-up credits are shared by the whole workspace
  • a team seat costs the same as the matching individual plan

creative

  • a new creative home with your recent generations one tap away
  • use several images as input for models that support it, on web and mobile

speech 🔊

  • text to speech and speech to text in one place, on web and mobile. upload audio and get a transcript

better notes ✍️

  • notes now sync live across web and mobile, so edits from different devices merge instead of overwriting each other
  • export a note together with its images

live mode 🎙️

  • live mode responds faster and sounds more natural
  • share your screen in live mode on iOS and Android
  • live mode is more reliable now and reconnects on its own if the connection drops on web

models

  • added GPT 6 Luna, GPT 6 Sol and GPT 6 Astra
  • added Claude 5.5 Opus, Grok 4.7, Qwen 3.8 Max, Qwen 3.8 Flash and DeepSeek v4.1 Flash

a much better ui ✨

  • a big ui refresh across web and mobile: cleaner, faster to get around, and consistent on every device, so it looks and works the same wherever you open zemith
  • a redesigned composer: tools, skills and plugins you pick show up inline as chips, and your saved prompts are right there too
  • one switch in the sidebar to jump between chat and cloud
  • organise your chats with labels on web and mobile

mobile 📱

  • mobile now has almost everything the web has: cloud, plugins, skills, workspaces and the new agent
  • plugin marketplace, plus creating and editing your own skills (with the made by zemith starters)
  • push notifications for cloud tasks and approvals
  • labels for notes and projects
  • podcasts keep playing when you lock your phone, with lock screen controls

Gemini 3.8 Flash and zAI GLM 5.3

  • Added Gemini 3.8 Flash and zAI GLM 5.3 to the supported text models in Zemith

Added Gemini 3.7 Flash and Grok 4.6

  • Added Gemini 3.7 Flash and Grok 4.6

Qwen 3.7 Flash

  • Added Qwen 3.7 Flash

Opus 5

  • Added Opus 5 to the supported text models in Zemith

Gemini 3.6 Flash and Gemini 3.5 Flash Lite

  • Added Gemini 3.6 Flash and Gemini 3.5 Flash Lite to the supported text models in Zemith

Kimi K3

  • Added Kimi K3, a fronttier open weight model that rivals with GPT 5.6 Sol and Claude Fable 5 to the supported text models in Zemith

GPT 5.6

  • Added GPT 5.6 Luna, Terra and Sol frontier models to the supported text models in Zemith

xAI Grok 4.5

  • Added xAI Grok 4.5, a frontier model to the supported text models in Zemith

Workflow Fix

  • Added support to attach output node to LLM node in workflow
  • Added support for copy paste of LLM text in LLM node within workflow

Sonnet 5

  • Added Claude Sonnet 5 to the supported text models
  • Added Prune AI to the support image generation models

GLM 5.2

  • Added zAI GLM 5.2 to the supported text models

Bug Fix

  • Fix issue where project library chat not able to access uploaded sources.

Kimi K2.7 Code

  • Added Kimi K2.7 Code to the supported model list for text generation

Qwen 3.7 and GLM 5.1

  • Added Qwen 3.7 Plus Qwen 3.7 Max and zAI GLM 5.1 to the supported model list

Added MiniMax M3

  • Added MiniMax M3 to the supported model list

Added Claude 4.8 Opus

  • Added Anthropic latest Claude 4.8 Opus

Added Gemini 3.5 Flash

  • Added Google latest Gemini 3.5 Flash

Added SeedDance 2.0 and HappyHorse v1.0 video generation model

  • Added SeedDance 2.0 and HappyHorse v1.0 video generation model
  • Fixed issue where document upload fails

Added DeepSeek v4 Pro and DeepSeek v4 Flash

  • Added DeepSeek v4 Pro and DeepSeek v4 Flash to the supported model list