morphyDocs

Settings

Every section of Morphy's Settings: account and plan, plan usage, activity, personalization, shortcuts, voice, skills, Morphy Agent and community.

Open Settings with Ctrl + ,, the gear at the bottom of the sidebar, or Ctrl + K → Settings. It reopens on the section you used last. Esc closes it.

Hover the ⓘ next to a setting (or tab to it) to see what it does.

General

  • Morphy and its updates: the version you're on, and Install updates automatically (on by default) under it. See Updates.
  • Agents:
    • Morphy runs on: the AI behind Morphy, your assistant.
    • New chats: how a new chat starts; chats you already have keep their own settings. The button shows the AI and its model, for example Claude · Default. Its menu works like a chat's model menu:
      • The models and Effort: the model and effort new chats of that AI start on. Until you pick others, that's Default (the AI's own default model) and Auto (effort left to the model). You can also set them from a chat with Use for new … chats at the bottom of its model menu.
      • Permissions: the permission mode new chats get: No guardrails, Auto-accept edits, Ask first or Plan only. Claude's Auto is picked per chat, in its model menu. A new install (from Morphy 0.1.22) starts with Auto-accept edits; an install from before keeps what it had.
      • Opens as: Chat (the default) or Terminal: new chats of that AI open as its own app in the card, like Claude Code or Codex (see Open a chat as a terminal). Only AIs that have an app of their own show it. With Terminal, the button shows a small terminal icon. Shift + double-click on the canvas makes the other kind.
      • Provider: which AI a double-click on the canvas starts. It also decides what New chat buttons and the canvas menu (right-click the canvas) make.
    • Open folders in (only with more than one code editor installed): which editor Open in … uses, from a chat's folder, a project's menu and Ctrl+K. Without a choice, it's the first of VS Code, Cursor, Windsurf, Antigravity IDE, Zed and VSCodium that Morphy finds.
    • Always ask before: commands agents ask you about first, even in No guardrails (Codex asks only in No guardrails, and Antigravity only in No guardrails on Windows for now). Empty until you add one. Type a command and press Enter (or Add); × removes one. See Always ask before.
    • Keep computer awake: stops your computer from going to sleep on its own, so agents don't freeze halfway through a task. Off (the default) leaves it to your computer's own sleep settings, While agents work lets it sleep again once they're done (work an agent left running in the background, like a dev server, counts), Always keeps it awake while Morphy is open. The screen can still turn off, and closing a laptop's lid still puts it to sleep. On battery, some Windows laptops still go to sleep a few minutes after their own sleep time.
    • Keep processor cores free (Windows only): agents always run at lower priority so your own work comes first. With this on, they also stay off a third of your processor cores while they build or render. Some antivirus programs dislike how Windows is asked to do that, so it's off by default.
    • Folders on other computers: lets a Claude or Codex chat work in a folder on another computer you reach with SSH, like a Raspberry Pi or a server. A chat's folder menu then has Use a folder on another computer… (see A folder on another computer). Off by default; turning it off keeps chats already set up working.
    • Start Morphy when I log in: opens Morphy for you when you sign in to this computer. Off by default. Your computer still has to be on and signed in: Morphy doesn't run as a background service, and nothing it does leaves your computer. Turning Morphy off in your computer's own list of startup apps (Task Manager on Windows, Login Items on a Mac, your desktop's startup applications on Linux) turns this off too. A Mac may ask you to allow it first: System Settings → General → Login Items.
    • Keep running when the window closes (Windows): closing Morphy's window leaves it running in the tray by the clock, so your chats and agents keep working. Click the icon to open Morphy again; right-click it for Open Morphy, Restart to update (when an update is ready) and Quit Morphy. Off, closing the window quits Morphy, asking first while agents work. The first time you close the window, Morphy asks which you want, with a box to also start Morphy when Windows starts. On a Mac, Morphy already keeps running when its window closes. Updates install when you quit.
    • Shipped work and suggestions (off by default): every agent gets one short extra instruction and ends each reply with a hidden line saying what it finished and what you'll most likely ask next. What it finished fills Work shipped, Recently shipped and Top projects in Activity. What you'll ask next shows greyed out in the chat's empty box; press Tab to put it in the box (see Send a message). It has almost no effect on your agents and uses barely any of your plan, and nothing goes anywhere but your own AI and your computer. Off, neither is asked for: what's already listed in Activity stays, but nothing new is added until you turn it on.
  • Workspace: Export… saves your projects, tabs, chats and cards as a zip, with each chat's latest messages if you like, and Import… adds one from another computer beside what you have. See Move to another computer.
  • Privacy and data: your Data folder (where everything is stored; see Your files), and Send error reports (off until you say yes). When something breaks, like an agent that stops, a turn that fails or a crash, Morphy sends what went wrong so it can be fixed: the error, the AI it happened with, the Morphy version, your system and your account ID. Never your chats, prompts, files or code, and your name and folder names are taken out first. Network trouble and plan limits aren't sent. The first time something breaks, Morphy asks at the bottom of the canvas: Send reports, No thanks, or × to be asked again next time. For your first two weeks there's also Share your first two weeks (on): Morphy tells its server when you first reach each step of getting started, like a setup screen, connecting an AI, your first project, message and reply, or the kind of problem when a turn fails, so we can see where new users get stuck. Only the steps and their times, with your account ID; never your chats, prompts, files, folder or project names. It stops by itself after two weeks, and the row goes with it.
  • Help shape Morphy: Send feedback opens the feedback form (see Report a bug, share an idea or write a review), and Discord opens the Morphy Discord.

Account

Your Morphy account (the Google or GitHub account you signed in with), with Sign out to switch to another account (also under your name at the bottom of the sidebar): Morphy restarts and asks you to sign in again. Your canvas, chats and AI connections stay on this computer, so signing out doesn't lock anything away; anyone who signs in next sees the same canvas. Click your name to type the one Morphy shows for you instead: in this card, at the bottom of the sidebar, in Morphy's greeting and in Morphy's notes about you (unless you changed the name there yourself). Enter saves it, Esc keeps the old one, and an empty name goes back to your account's. Click your picture to show your initial instead (click it again for the picture). Both stay inside Morphy on this computer: your account, Morphy's emails and the leaderboard keep their own names. Under it, Discord: Connect opens Discord in your browser; click Authorize there and Morphy adds you to the Morphy Discord server with the Verified role and shows your Discord name and picture here. It gets your Discord name and picture, not your email. Disconnect takes the role back (you stay in the server). An account from before this feature first confirms itself with Google in the browser, once. In the same card, Show Morphy on my Discord profile (off until you turn it on): while Morphy is open, your Discord profile says Playing Morphy, with how many agents are working, the models they run on, how many tokens they wrote today, how long Morphy has been open and a Get Morphy button. It never shows your projects, files or chats. It needs the Discord desktop app open on the same computer (not Discord in a browser), but not a connected Discord account; the line under the switch says whether it's showing. Then Plan: Freemium or Pro, Pro's price, Check again and Get Pro (see Freemium and Pro); Pro for life has a gift at its end, Support Morphy, which opens the supporter page in your browser: every month or once (once you've supported, the card thanks you instead, with Manage for a monthly support). Then Connections: one card per AI with its status and what you can do, like Connect, Reconnect, Sign out or Check again. The Claude Code, Codex and Grok cards can hold more than one account: Add another account, then Switch (see More than one account). With two or more, Advanced under the list has Switch accounts when a limit is hit (off: the chat offers it) and which account each project keeps; a single chat picks its own from its model button's Account (see limits and projects). At the bottom, Workspace keys (.env) opens your API keys file as a card. See Connect your AI.

Plan usage

How much of your Claude, Codex, Grok and Antigravity plans you've used, and when each limit resets. Only the AIs you've connected show, the one you used last first (or one close to its limit). Antigravity shows one limit per group of models that share it, like Gemini models and Claude and GPT models. Grok shows This week, and This month too if your xAI plan has a monthly limit. Grok's numbers come with the sign-in its CLI keeps, which Grok renews while it works: if you haven't used Grok for a few hours, Morphy shows the last numbers until your next Grok chat. With only an xAI API key, Grok is billed per use and has no plan limits to show. When OpenAI gives your Codex account a free limit reset, Codex's card says so under its limits, with Use it. A reset can't be undone, so Morphy asks first: Use resets your weekly and 5-hour limits right away and the numbers update. If it doesn't work, the line says why, with Try again. Refresh updates it. You can also see it from the canvas: the bottom-left corner shows the two that matter most, and selecting a chat and clicking Plan usage in the toolbar shows them all. Morphy checks at most once a minute; when an AI doesn't answer for a while, it keeps showing the last numbers, says how old they are, and tries again by itself. The last numbers stay after a restart too, so they show from the moment Morphy opens. A limit whose reset time has passed shows as reset. While Claude works, its numbers follow every reply, so they stay current even when Claude's usage check is busy (another app reading it with the same login, like a usage monitor, can keep it busy). Copilot, OpenCode, OpenRouter and Ollama don't report plan limits to Morphy.

Activity

Your agents' work over time:

  • The numbers at the top: your streak, today, the last 7 days and your prompts in total, then all time: tokens, lines of code, how long your agents worked and on how many days you used Morphy.
  • Daily activity: prompts, tokens and lines of code over 14, 30 or 90 days, with a line for the 7-day average. The ⓘ beside it says what each one counts.
  • All time: your total tokens, prompts or lines of code, added up day by day, so you see it grow.
  • Rhythm: when in the week you work.
  • When you work: your prompts by hour of the day and by day of the week.
  • By AI: how much of your prompts, tokens and agent time went to each AI.
  • Records: your best day, week and month so far, your longest streak and the longest an agent worked on one message.
  • Work shipped: what your agents finished, by kind of work or by model, over 14, 30 or 90 days or all time, with how much agent time it took.
  • Recently shipped: the latest finished work, with its project.
  • Top projects: the projects your agents finished the most work in.

Point at a day, a bar, a number or a record to see more about it in a small card: for a day, its prompts, tokens, lines of code and agent time, which AIs you used and your busiest hour. On a line, the card follows your pointer from day to day. Everything here is counted on your computer and stays there. Hours, AIs and agent time are new: they're counted from the day you update, so those cards fill up as you go.

On Freemium the charts show the last 14 days; the longer ranges, older weeks and older days in detail come with Pro.

The three shipped-work cards fill up while Shipped work and suggestions is on, in General.

Personalization

The Color scheme (System, Light or Dark), an Accent colour, Interface (interface size, density, card style and chat text), Morphy (the little Morphy on the chats: Classic, Crystal, Pixel, Ink, Cloud, Slime, Flame, Holo or Clay, its size, on the box or not), the Canvas grid, Motion, the canvas Background (built-in wallpapers, animated shaders and your own images or videos, and how much a picture or video is dimmed), your Favourite background, and the Sound a chat's bell plays: a built-in or a file of your own. See Personalization.

Shortcuts

Every shortcut and mouse gesture, with a search box. Click a shortcut to change it. See Keyboard shortcuts.

Voice

Voice typing: download the Speech model (670 MB, once), turn Voice typing on or off, pick the Microphone, see the Shortcut, choose whether to Show words while you talk and Send when you stop talking, turn on Clean up with AI (Pro), and Try it. See Voice typing.

Morphy Agent

Which AI runs Morphy (Change takes you to General → Morphy runs on), the three instruction files Morphy reads before every message and its folder (Instructions), the tools it can use (Connections: GitHub, Notion, Linear, Sentry, Context7, Supabase, Cloudflare, Stripe, Apollo.io, vidIQ, Higgsfield, fal, ElevenLabs, Dropbox, Metricool, Wix, Figma and Obsidian, each with Connect right in the list, most signed in through your browser, and Missing a service? at the end to ask for one; a connected one opens its own page, where Use in gives it to Morphy only, the chats of the projects you pick, or every chat; an MCP server of your own is in Ctrl+K; see Connect your tools), and the keys it can reach (Keys → Morphy's .env file, which opens to its keys by service, with All variables and Import… to add a batch). See Morphy, your assistant.

Skills

Your library of skills: create, edit, import and delete them.

Community

What people share.

  • Leaderboard: how much your AI wrote today, this week and this month, and the board of everyone who joined. Joining needs Discord connected. See Leaderboard.
  • Apps: bigger than cards, each living where you put it. Your apps lists the ones you have, each with a switch to turn it on or off and its version. Available lists what you can install, such as the Morphy cockpit (part of Pro): Install downloads it, checks it and opens it. Ctrl+K finds this page as Apps.
  • Playground: the libraries people share, one card each, with your Morphy in its picture, in the style you picked under Personalization. Click a card to open it, and Back to return.
    • Content skills: free formats for short videos, each a SKILL.md your AI follows. Put one on a chat, add it to your skills or download it; the line at its foot opens our Discord, for more skills and repos. See Content skills and flow templates.
    • Flow templates: flows to build in one click: ours for building software, and the ones people shared. Save or export your own from a flow's ⋯ menu, and share it in our Discord. See Flow templates.
    • Cards library: cards other people made, ready to put on your canvas. See Cards.

On this page