What's new in Mnevra.

A native macOS app, sharing one account and one graph with your phone — and a few desktop-shaped extras.

• A proper desktop app. All of Mnevra you already use — chat, the knowledge graph, biographies, todos, goals, reminders, knowledge review, spaces — built for macOS instead of wedged into a window.
• Menu-bar quick capture. A small popover lives in your menu bar. Type something worth keeping and Mnevra files it in the background — no full window needed.
• Menu-bar todos panel. Peek at what's on Now, Next, Later or Done from the menu bar. Each state has its own inline composer, you can drag between sections, and one click takes you through to the full list.
• Selection → todo, from any app. Highlight text anywhere on the Mac, choose Mnevra from the macOS Services menu, and it lands as a draft todo with the selection as notes. AI suggestions help shape a sensible title and pick the right list.
• Open at login. Set Mnevra to launch quietly when you sign in, so the menu bar is ready before you reach for it.
• App lock that fits the Mac. Locks on cold launch, system or screen sleep, fast user switching, and a Control-Command-L menu shortcut — but doesn't re-lock every time you Cmd-Tab away.
• Push notifications on macOS. Reminders, nudges and digests reach you on the Mac too, with the same reply-from-the-banner behaviour as iOS.
• Cross-space engagement nudges. When Mnevra nudges you about a person, the conversation opens scoped to whichever space that person lives in — even if you're currently in a different one.

Voice mode and the share extension stay iPhone-only for now.

• API tokens. A new Settings screen for minting scoped Sanctum tokens you can hand to other tools. Name them, see when they were last used, revoke any time.
• MCP server. A small Model Context Protocol server lets Claude Code or Claude Desktop read and mutate your todos directly — list, add, complete, update, delete — using a token you control. Pipe ideas into Mnevra from wherever you already write.

A run of smaller updates that didn't earn their own headline, but add up.

• Real-time sync between devices. Add a memory on your phone and it appears on your Mac without a refresh — and vice versa. Silent push and live broadcast channels keep both apps in step.
• "Today" view for todos. A dedicated lens showing only what you've committed to right now, across every list. Reorder with a drag to set your own priority.
• Quick capture, but conversational. If a quick capture is ambiguous — which list, what date, who is "her" — Mnevra can ask one short follow-up before filing, instead of guessing.
• Goals can wait on someone else. Mark a goal or step as waiting on a third party and Mnevra stops pushing you about it, but keeps a gentle chase-up cadence so it doesn't drop off the radar.
• Snooze a goal. Mute proactive nudges on a goal or a specific step until you're ready to come back to it.
• Entity avatars. Upload a photo for any person, pet or place — or let Mnevra pull a likely portrait from the photos already in your graph.
• A few more agent tools. Mnevra can now delete reminders for you, set or clear an entity's avatar, and reach into other spaces when a nudge is about a person who lives in a different one.
• Tidier biography interviewer. Chapter chats keep their gap list stable across runs, work through the backlog properly, and end with a clean wrap-up instead of trailing off.

Mnevra can now hold the short-term stuff too — the calls, errands and follow-ups that sit between "remember this" and "remind me".

• Todos, properly. Create tasks with a title, notes and an optional due date. Each one lives in a list you pick, and you can have as many lists per space as you want (Errands, Follow-ups, Reading, whatever suits).
• Now, Next, Later, Done. Four simple states instead of a sea of checkboxes. "Next" is your default queue, "Now" is what you're actually doing today, "Later" is the backburner, and "Done" gets out of your way.
• Promote to today. One tap on anything in Next or Later moves it into Now, with a timestamp so you can see when you committed to it.
• Linked to the people they're about. Attach a todo to a person, pet or place and it shows up on that entity's profile. Open Mum's profile and you'll see "call about the house" sitting there waiting.
• Linked to reminders. Turn any todo into a scheduled push with a time and timezone — or attach an existing reminder to a todo so the two stay in sync.
• Linked to goals. Break a goal into concrete todos instead of abstract steps. Each one can be ticked off independently and feeds into your goal progress.
• Ask Mnevra to manage it for you. "Add a todo to book the dentist next week", "what's on my list today?", "mark the passport renewal done" — the agent can create, update, complete, promote and search todos directly.
• Per-space, like everything else. Your Work todos stay in Work. Your Personal todos stay in Personal. Nothing leaks between spaces.
• On the home screen. Todos in the "Now" state surface on the home screen for the space you're in, so the next thing is always one tap away.

• No more repeating-question loops. When the app asks follow-up questions about an import, it now remembers what you already answered and won't come back with the same (or very similar) question again. Clarifications are capped at two rounds, after which the app commits to its best interpretation rather than stalling.
• Imports land in the right space. Every imported file now records its memories, people and places against the space you were in when you uploaded it — no more bleed between your Personal and Work life.
• Friendlier errors. If something goes wrong during an import you'll see a clear sentence instead of a raw database error.

The big one. Mnevra can now hold more than one side of your life at once without mixing them up.

• Create a space for anything. Personal is still there by default, but you can now spin up separate spaces for Work, Health, a specific project, or anything else you want to keep its own set of people, places and memories.
• Templates to start you off. Pick a template (Work, Health, Project or Custom) when you create a space and Mnevra preconfigures a sensible tone, icon, colour, check-in cadence and what shows up on the home screen. You can still tweak anything afterwards.
• AI-assisted setup. Describe what the space is for in plain English and Mnevra will draft a name, icon, colour, custom instructions and starter prompts for you. If you'd rather skip that, the template defaults are a solid starting point.
• Per-space personality. Each space has its own icon, colour and custom instructions that guide how the AI talks to you there — professional in Work, supportive in Health, whatever suits.
• Per-space home screen. The greeting, the "what could you tell me about…" prompts and the suggestion chips are all tailored to the space you're in. You can regenerate them at any time from the space's settings.
• Your Work data stays in Work. People, places and relationships you add inside a space are invisible from every other space. The only thing shared is you — your own identity stays consistent wherever you are.
• Switch from anywhere. The space switcher lives behind the side-menu button. Your icon and colour in the chat header tell you at a glance which space you're currently in.
• Interface colour follows the space. The whole app — links, selected chats, buttons — picks up the accent colour of the space you're in, so context is never ambiguous.

The morning digest has been cleaned up so it actually reflects what's going on in your life.

• No more messages about people who've passed away. Birthdays, anniversaries and check-in prompts now respect the died / death_date info on entities. Nobody wants a cheerful reminder about a late friend.
• Just-mentioned people don't come back the next morning. Mentioning someone in chat now counts as a check-in, so they won't be in tomorrow's "you haven't spoken about…" list.
• Only notify once per person per day. You won't get the same "check in with so-and-so" notification twice in the same day.
• More prompts on the home screen. Gaps in your biography (like "where were you born?") now reliably show up as suggestion chips instead of getting crowded out.

• Check-in cadences. Every person Mnevra knows about can now have a check-in rhythm — weekly, monthly, quarterly, whatever fits. When it's been too long, you'll get a gentle nudge.
• Batch cadence setup. From the Knowledge Review screen you can set check-in cadences for lots of people at once, with Mnevra suggesting sensible defaults based on how close you are.
• Goal steps. Break goals into concrete steps and check them off as you go, with a running sense of progress.
• Search inside your conversations. A search bar on the chat list lets you jump straight to past conversations by keyword. The list also paginates properly for long histories.

• Push notifications that can be replied to in-place. When Mnevra nudges you, you can tap reply on the notification and keep the conversation going without even opening the app fully.
• A proper notifications inbox. Every push Mnevra has sent you lives in a dedicated list so you can catch up on anything you missed.
• Unread badges in the side menu. The hamburger button shows a count of unread notifications when the side panel is collapsed.
• Face ID app lock. Lock Mnevra behind Face ID (or your device passcode) so your memories stay private even on an unlocked phone.
• Timeline view. See your life laid out chronologically, with key events grouped by year.
• Smarter entity profiles. Tap an entity in chat and you get a richer profile sheet with their memories, relationships, photos and notes.

• Better biographies. Chapters are now generated with a more structured approach — a dedicated pass for photo placement, better captions, proper text wrapping around images, and cover generation that actually looks like a cover.
• Share extension. Select any text in Safari, Mail, Notes or anywhere else, tap Share, and send it straight to Mnevra as a quick capture — no copy/paste needed.
• EXIF-aware photos. When you add a photo, Mnevra now reads its embedded date and GPS and tags the resulting memory with the right time and place automatically.
• Replace a photo in-place. You can now update an existing photo (caption, crop, replace the file) without deleting and re-adding it.
• Thumbnails everywhere. Image lists are noticeably faster thanks to pre-generated thumbnails.

Paste a URL or drop in a file (PDF, text, markdown, Word doc) and Mnevra will pull the people, places and memories out of it.

• Big files handled properly. Large imports now use Anthropic's batch API under the hood so they don't time out.
• Duplicate detection. The review screen tells you if a person or place it extracted is probably one you already have, so you don't end up with two of everything.
• Delete imports. Change your mind about a file you uploaded? You can now delete the import and its preview outright.

• A proper home screen. When you open a fresh chat you now get a greeting, a snapshot of what Mnevra knows, and a set of context-aware suggestions to start a conversation.
• Knowledge Review gets smarter. The review screen now flags a much wider set of data-quality issues — overlong summaries, subtle name variants, conflicting dates — on top of the duplicates it already caught.
• Export your data. You can now export everything Mnevra holds for you in structured form, at any time.

• Private memories. Mark any memory as private and Mnevra will still use it in conversation, but keep it out of your biography and any exports.
• Biography exclusions. Tell Mnevra what not to include in the biography — specific topics, people or stories — and it respects those rules on every regeneration. You can manage the exclusion list from the biography screen.
• Memoir export mode. Export your biography as a memoir instead of a straight biography for a more personal narrative voice.
• First-run onboarding. New accounts now get a guided setup that introduces what Mnevra does and walks you through connecting your API keys.
• Bring your own Anthropic and OpenAI keys. Prefer to handle usage and billing directly? You can now plug in your own keys from Settings, and usage is logged against them.
• Account management. Edit your name, email, password and avatar all from one screen.
• Timezone-aware reminders. Reminders and check-ins now fire at the right local time wherever you are.
• Privacy at the database layer. Every read is now scoped to your own account at the database level, so even a bug in the app can't surface someone else's data.

• Reminders. Ask Mnevra to remind you about something — a call, an anniversary, a task — and it'll push a notification at the right time.
• Entity deep-dive. Tap a person or place in chat to start a conversation focused entirely on them. Great for catching up on what Mnevra knows, or filling in missing details.
• Reply to a push. Tap reply on a Mnevra push notification to continue the conversation without opening the app fully.

A big week that turned Mnevra from a chat into a proper life-story tool.

• Your biography, generated on demand. Mnevra can assemble what it knows about you into a structured biography — chapters, prose and illustrations pulled from your own photos.
• Interview mode. Chapters with gaps trigger a guided interview where Mnevra asks a direct opening question and fills in the missing beats as you answer.
• PDF export, three designs. Export your biography as a printable PDF in Modern, Classic or Magazine style, emailed to you when it's ready.
• Highlight and correct. Spot something wrong in the biography? Highlight it and chat with Mnevra to fix it — corrections become directives it remembers on every future regeneration.
• Photos in chat. Attach photos to a message and Mnevra stores them against the relevant people and memories, with a full-screen viewer when you tap them.
• Push notifications. Mnevra can now reach out over push (the foundation for reminders and the daily digest that come later).
• Forgot password. Reset your password from the login screen without contacting support.
• Stream recovery. If a reply fails mid-stream the app retries it automatically instead of leaving the conversation stuck.
• Knowledge Review arrives. A first version of the review screen that surfaces obvious duplicates and overlapping memories so you can merge them in one tap.

• Quick Note. Jot something down without opening a full conversation. Mnevra processes it in the background and files the people, places and memories into the right slots automatically.

• Inline thinking. When Mnevra is working on something complex, its reasoning unfolds inline in the chat so you can see what it's considering.
• Avatar upload. Add or replace your own profile photo, with in-app cropping.
• Browse everything Mnevra knows. A searchable list of every person and place in your knowledge graph.
• Steadier streaming. A failing tool call no longer crashes a reply mid-stream — errors surface as a message and the conversation keeps going.
• Timeline fixes. Memories spanning multiple dates are split correctly so your timeline stays accurate.

The first build. Mnevra opens as a chat and quietly learns about your life as you talk.

• A conversation that remembers. Talk to Mnevra about anything — people, places, moments — and it builds a persistent picture of your life in the background.
• People, places and memories. Everything you mention gets captured as a person or place entity, with memories attached and relationships mapped between them.
• Voice mode. Tap into a realtime voice conversation with Mnevra when you don't feel like typing.
• Multiple conversations. Start a new chat any time. Old conversations live in the side panel with auto-generated titles so you can find them again.
• Goals. Track personal goals alongside the conversation that created them.
• Personas. Pick how Mnevra talks to you — warmer, sharper, more direct — and switch any time.
• iPhone app. Native iOS app with account creation, login and profile.