Organize your pages in databases
Create databases in Pages to organize journals, reports and logs. Add text, number, date, people, selection and checkbox columns, then search, filter and sort your entries. Each entry opens as a full page. Numo and MCP-connected agents can read and update your databases too.
Connect your services to Numo
Add a personal MCP connection in your account settings so Numo can use your other services in chats and coding tasks. Scheduled routines use the project owner's connections.
Ask Numo to work in another project
Create or update an issue in another project without leaving your conversation. Tell Numo which project to use; otherwise, it keeps working in the current one.
Give your routines skills and project context
Mention issues, pages or objectives in a routine's instructions, and select repository skills to guide the work. The agent receives that context whenever the routine runs.
Give Numo your project's reusable instructions
Select a skill from your repository when writing to Numo. You can preview its instructions before sending them and use skills from a local checkout in the desktop app too.
See whether an objective is on track
Each objective now shows its recent pace and eight weeks of completed issues. When there is enough activity, an estimated finish date helps you spot a deadline at risk.
See what is holding up a pull request
The pull request view now lists missing approvals, unresolved discussions and checks still to pass. Update the branch or rerun failed checks from minddy when your repository allows it.
Resume an agent after a failed run
When a failed run has saved progress, you can send Numo a follow-up and continue in the same conversation. The work already done remains available.
Highlight important notes in your pages
Add a colored callout to draw attention to a decision, tip or warning. Choose its color and icon, or paste an existing callout from Notion.
Ask Numo about a passage in a page
Mention Numo in a page comment to ask a question about the selected passage. Its reply stays in the same thread, beside the text you are discussing.
Preview attachments without leaving minddy
Open images, PDFs, text files, audio and video in a large viewer. You can read, watch or listen without downloading each file first.
Use minddy on macOS, Windows and Linux
minddy now has its own desktop app on all three platforms, with automatic updates and a server picker for Cloud or self-hosted instances. Download the build made for your system and work outside the browser.
Take stock of what you have accomplished
Your personal statistics now give you a clearer view of completed issues, pace and time spent. Breakdowns by project, category and objective help you understand what you have accomplished across all your work.
Choose the face you show in minddy
Pick a generated Lorelei avatar or upload your own image from your profile. The same avatar follows you through projects, comments and conversations.
Add documentation to your public feedback page
Published project pages can now appear as tabs on your public feedback page. Add a roadmap, documentation or community rules for visitors to read alongside suggestions.
Move your account to another minddy instance
Export your account, then import it into another minddy instance. Projects, issues, pages, settings and personal history come with you; passwords, provider keys and subscriptions stay behind.
Use your own AI provider across minddy
One key can now power Numo, Smart-fill, transcription, feedback review and the rest of minddy’s AI. Compatible local models work too, so supported tasks can stay on your computer.
Keep GitHub issue dependencies in sync
When a GitHub issue blocks another one, the linked minddy issues now show the same relationship. Changes made on GitHub continue to update the issue panel automatically.
See page edits as they happen
When someone else edits a project page, their changes now appear while you have it open. Autosave keeps your own work in step instead of leaving two tabs to overwrite each other.
Run the coding agent on your own computer
The desktop app can attach a project to a local folder and run Numo there. You see file changes as it works, and they stay in your working copy until you decide what to do with them.
Keep your project's notes and specs in pages
Create project documents and organize them into pages and subpages in the sidebar. Add text blocks and notebook tasks, then link a page to an issue as a resource. Numo, minddy’s AI assistant, can read and edit these pages too.
AI suggests details for your new issues
Smart-fill uses AI to read an issue’s title and description and fill in its priority, effort, categories and objective. You can turn it off for one issue or for all issues in your account settings.
Select several cards on the board and move them together
Draw a rectangle across the board to select the cards it covers. Dragging one of them carries the whole selection to another column, in the order you see them.
Repository issues and minddy issues stay in step
Issues imported from GitHub or GitLab arrive with their labels, their priority and their assignee, and keep following the remote one. Closing an issue in minddy now closes it on the repository too.
Invite someone who doesn't have an account yet
Add them by email address, and they get an invitation to join the project. They create their account from that email and land straight in the project. Pending invitations expire on their own.
Schedule recurring work for your coding agent
Give the coding agent instructions for a repository and choose a schedule, such as a security review every Monday. Each routine runs automatically, proposes fixes in a pull request and reports its findings. Review past runs in the Routines tab.
Get a notification when something lands in your inbox
Notifications turn on device by device, from the device itself. Each one is listed in your account settings, with when it was added and its last notification, and can be switched off or removed. On iPhone, add minddy to your home screen first.
Start a project from a brief, and get your first issues
Creating a project now asks where you are starting from. Coming from another tool, you drop your CSV export. Starting from nothing, you describe the project in a few lines: Numo asks what is missing, then proposes the issues and objectives, which you review before they exist.
Issues find their assignee on their own
Describe in plain words the kind of work each member takes on, and any issue leaving triage without an assignee gets one. It switches on in the project settings. Alone on a project, the issues simply come to you.
Your feedback board now takes comments
Visitors can now comment on a suggestion and reply to other comments on your public feedback page. They verify their email before their first comment; only an avatar is shown beside it. You can disable comments at any time.
Attach a link to an issue, not just a file
Issues and objectives now hold resources: files as before, and web addresses. Drop a file straight onto an issue to attach it. Everything sits in one place, ready to open or download.
Add a feedback button to your app with guided setup
Choose how to collect suggestions, whether users are already signed in and where the button should appear in your app. The guide prepares instructions for your coding assistant. You can also give them to Numo to implement the change and open a pull request.
Feedback written in another language arrives translated
Numo, which already reads every submission before it goes live, now translates the ones written in a language other than yours. The original stays one click away. It works on feedback from the board and from your own app alike.
Export your issues as a CSV file
Pick the projects and the statuses to include, and get a file. It follows minddy's own import format, so it can be dropped straight back into another project.
Review and merge your pull requests without leaving minddy
Every pull request on a linked repository now opens inside minddy: the changes line by line, the commits, the checks and the whole conversation. You comment, approve, resolve a thread or merge, under your own GitHub or GitLab account.
Ask Numo to review a pull request
Numo reads the changes, the issue and the discussion, then posts a summary and its comments on the lines concerned. You reply to them like any other review. And when you are the one requesting changes, Numo can go back and make them.
Let the agent chain plan, code and checks on its own
Pick an automation preset, the projects and the efforts it runs on, and an issue reaching To do starts the loop by itself: the plan, the implementation, then a check of the work. On the biggest ones, it waits for your go-ahead before writing any code.
Issues that come back on their own
An issue can now repeat every day, week, month or year. Close it, and the next one is already waiting in the backlog, at the right date. The project settings list every recurring issue, to change a rhythm or stop one.
Run the coding agent on your own API key
Add an Anthropic, OpenAI, Google, OpenRouter or compatible key, at signup or later from your account. The agent's runs are uncapped; your provider bills them on whichever model you want. The key is encrypted and removable in one click.
Mention a person, an issue or an objective as you write
In a description or a comment, typing an at sign now suggests the people on the project, its issues, its objectives and your projects. What you pick becomes a link others can follow, and anyone you name is notified.
Deleted issues wait in the trash
Issues, objectives, feedback and projects you delete land in the trash for 30 days, with who removed them and when. Restore one in a click, or empty the trash yourself.
Bring in the issues you already have
minddy reads a Linear, Jira or Trello CSV export, at signup or later from the project settings. And when a project is linked to a GitHub or GitLab repository, every issue opened there arrives here, its status following the remote one.
The home page shows what needs you now
The home page now lists issues with approaching deadlines, with overdue work first. A second section brings together incoming issues and feedback waiting for review across your projects. Each item shows which project it belongs to.
Keep the board, hide the rest
Zen mode puts away the sidebar and the header, leaving your work alone on screen. It turns on from the command palette, and off the same way.
Take everything with you, or delete it all
The account settings now export a JSON file holding your account, projects, issues, comments and conversations. The same page deletes your account for good, after telling you exactly what goes with it.
Brief the agent before it starts
Launching the coding agent on an issue now takes an instruction of your own, on top of what the issue already says. You also pick how long it thinks before acting, from a short pass to the deepest reasoning.
A setup guide for connecting AI assistants
Connect an AI assistant to read and update your minddy projects through MCP. The new guide provides the server address, setup instructions for each supported assistant and examples to try. It also includes instructions you can copy to your assistant to help set up the connection.
The public site now speaks French at its own address
Pricing, legal pages and the home page each exist at a French URL of their own, instead of switching language behind one address. The pages also got noticeably lighter, so they open faster on a phone.
Search issues and objectives from anywhere
The command palette now searches every issue and objective you have access to, across projects, and no longer just the ones on screen. Results come back as you type.
Launch an agent from the task notebook
Ask the coding agent to work on a task from your personal notebook. You can launch it directly from the task, without creating a project issue first.
Notifications you can actually keep up with
Notifications are now grouped by issue so you can follow a conversation without reading each event separately. Choose which types of activity appear in your inbox from your notification settings.
Create a project, add your first tasks, and invite your team. Connect an AI agent when you want help with the work.