AI Now

AI Now is the place where you work directly with your saved knowledge. It runs with your Mem server, so the desktop app, web app, and mobile clients can see the same AI Now sessions when they connect to that server.

Instead of starting from a blank prompt every time, AI Now starts from what you already know. That makes it useful for research, analysis, presentations, and longer multi-step tasks that should build on your own history rather than generic model output.

The First Useful Task

If you are opening AI Now for the first time, start with something grounded in your own knowledge:

What architecture decisions have I made about caching?

That is a better first task than a generic prompt because it lets you see the product model immediately:

• AI Now reads from your saved knowledge
• it can pull in files or enabled plugins when relevant
• it can turn the result back into updated knowledge if needed

What AI Now Can Do

Files, web research, plugins, and schedules need backend support. See Using AI Now from another device for these limits. You do not need extra skills or plugins for your first question.

Getting Started

Choose a task model and AI identity

Use the task's Model menu to select a configured provider and model, or Settings default. Wait for the current turn to finish before switching. The choice applies from the next turn to this task only; it does not change background processing or other tasks' defaults.

If you have created AI identities, select one before creating a task. The selection applies to the next new task; a resumed task keeps its original identity. Identity and Space are separate choices. Selecting an identity does not grant extra access to data.

Mention memories in your chat

Type @ to search and select memories, conversations, or Library content in the current Space. Check the composer references before sending. Request live note-plugin queries directly in the task; enabling a plugin does not automatically add external notes to this list.

Scheduled tasks

Use a scheduled task for work that repeats, such as a daily work record or a weekly document review. Each run uses your saved instruction and keeps its own result.

Create and activate a task

1. Open AI Now → Schedules, then click New schedule.
2. Enter a Name and What should AI Now do? Describe what to read, where to save the result, and how to verify it. Each run starts with fresh context; it does not continue the original chat.
3. Under When, choose Every day, Weekdays, Selected days, Every week, or Every month. Set the time and Time zone, for example 23:00 and Asia/Tokyo. The task follows the local clock in that time zone, including daylight-saving changes. Check the preview of future run times.
4. Choose Execution permissions and the policy for If Mem was offline.
5. Click Create draft, review the task, then click Review & activate. Check that it shows Active and a Next run time.

You can also ask AI Now to create a draft in chat: "Create a daily work-record schedule for 23:00 in Asia/Tokyo." Review and activate it in Schedules. A one-time reminder in chat does not create a recurring task.

Execution permissions

Enable Nowledge to allow the task to use these knowledge resources within its approved scope:

Knowledge writes stay in the task's Space. Schedules created in the App use the Default Space. Changing the Space in a chat does not change an existing schedule's destination.

A new Skill still needs human review and activation. An update prepares a version for review; it does not change the active content. Open Skills to inspect and apply the change. The allowed existing Skills are fixed when you save the permission grant. New Skills do not automatically gain update permission.

Web research allows searches and reads of public web pages. Nowledge and Web research do not allow authenticated browser sessions, shell commands, or arbitrary file operations. For an external service, use External connector actions to review an installed HTTP connector's exact action, account, and destination scope, then confirm the task. Enabling a plugin or choosing Auto in a chat does not grant these schedule permissions.

An App update does not add permissions to an existing task. If the editor says its resource permissions need an update, click Save changes, then Review & activate. Changing the instruction or execution permissions also pauses the task for a new review. A request outside the approved scope stops the run and asks you to review access.

Pause, edit, and handle offline time

Open a task and click Pause to stop future scheduling. You can also pause a running task, edit it immediately, and Save changes. An operation that already started is allowed to finish safely. Wait for it to end before activating or resuming the task. Pause does not undo saved changes; check recent runs and affected resources before resuming.

For a manual pause with no instruction or permission changes, click Resume. If you changed the instruction or permissions, lost a connection, or need to review a previous run, resolve the displayed reason and use Review & activate. To remove a schedule, click Delete and confirm.

Under If Mem was offline, choose:

• Run once when AI Now returns: combine missed occurrences into at most one run when the Mem backend becomes available.
• Skip missed runs: wait for the next future occurrence without running missed work.

A due task may wait for current background work to finish. It is not guaranteed to start at the exact scheduled time. Compare Scheduled for and Started at in the run record to check for a delay.

Check results and failures

Open the task and find Recent runs. Click a record to expand its full result, scheduled and actual start times, tool-call count, and error code when present.

An AI response alone does not prove completion. A useful read-only task can show Skipped. Repeated failures, invalid permissions, or an operation that may have started before shutdown can pause the schedule. Review its reason and confirm it again after resolving the problem.

Example: create a daily Memory and read it back

Create a task named "Schedule check". Select Every day, 23:00, and Asia/Tokyo. Enable Nowledge, leave Web research off, and choose Skip missed runs. Enter:

Use today's date in Asia/Tokyo. In the task's Space, find a Memory titled "Schedule check YYYY-MM-DD". If it does not exist, create it with the content "Schedule verified on YYYY-MM-DD". Then read that Memory by ID and verify its title, date, content, and Space. Report the actual ID and what you read. If today's Memory already exists, read and verify it without creating a duplicate. If a write or read fails, explain the failure and do not report completion.

Click Create draft, then Review & activate. To test it soon, set the time to a few minutes from now and check that the preview points to today. Keep the machine and Mem backend running.

After the run, expand Recent runs and check the Memory ID and read-back result. In Memories, select the Default Space, search for Schedule check, and open today's Memory to verify its content. The first successful creation and verification should show Completed. A later run on the same day can show Skipped if it verifies the existing Memory.

After testing, pause the task, change the time back to 23:00, save, and resume it, or delete the test schedule. A time change affects future runs; it does not repeat the write you just verified.

Deep Research

For comprehensive research, AI Now runs parallel sub-tasks across multiple sources and synthesizes the results.

How It Works

Ask a research question:

Research the current state of quantum error correction

AI Now will:

1. Search your memories for existing knowledge on the topic
2. Search the web from multiple angles
3. Synthesize findings into a single answer
4. Return findings with sources for you to check

Optional Skills

Skills are specialized capabilities you enable for specific tasks.

Enable skills in AI Now → Plugins → Skills.

Run and improve your own Skill

Type / in the composer, select an available Skill, add this task's goal and references, and send. Check the selected Skill first. Selecting it supplies the procedure for this run; it does not prove the result is complete.

You can ask AI Now to turn verified steps into a Skill or propose an update to an existing Skill. New content is a draft and updates are pending review. Review and activate or apply them in Skills before using them in future work. Generated text alone does not mean a Skill is active.

File Analysis

Attach files or folders to your conversation for analysis.

When you send the message, AI Now copies those attachments into the current task workspace under inputs/. The agent reads that workspace copy and writes new charts, scripts, reports, or slides back into the same task folder.

If you are connected to your Mem server through Access Anywhere, the App or Web client securely uploads the selected files into that task before AI Now starts. The server works from its own copy, so the original path does not need to exist on the server. If file attachment is unavailable for a hosted workspace, add the file to your Library and reference it from there.

You can also paste images; image analysis needs a model that supports images. Task attachments do not automatically appear in the Library. Save them there separately for long-term access.

Supported Files

Convert legacy Word .doc files to .docx before attaching them for analysis.

Example

1. Attach sales_q4.xlsx
2. Ask: "What are the top 3 trends in this data?"
3. AI Now analyzes and generates visualizations

Whole folders work too.

Save results and task files

After generating a report, chart, or presentation, use Browse task files to inspect the folder, preview files, or download them. Where the file menu offers Save to Library, save the file, then open its Source in the Library and check the content.

To save a conclusion as a Memory, ask:

Save the verified conclusion as a Memory with its sources and applicable date. Read it back by ID and report the actual Memory ID. Do not save unchecked guesses as facts.

A successful save should give you an actual resource ID, file, or source you can open. Check the resource if the only evidence is a text reply saying "saved." Deleting a task deletes its messages and working-directory files. Download or save outputs you need to keep first.

Presentations

Create a presentation based on our above study and research, include some charts or diagrams to support the insights

AI Now generates slides with structure, charts, and insights from your conversation.

Refine with follow-up requests ("Make the third slide more visual", "Add a slide about customer segments"), or click Edit to edit directly.

Click HTML to download a single .bento.html file and open it in a browser. You can also export to PowerPoint with the PPTX button; check layout, fonts, and charts after export.

Travel Planning

Plan a 5-day trip to Tokyo focusing on food and culture

AI Now generates an interactive day-by-day itinerary using your recent memories and web research as context.

Browse Now

Turn this on when a task needs your actual browser session, not a generic fetch. Typical cases:

• logged-in websites
• clicking, typing, scrolling, or multi-step flows
• screenshots or rendered-page extraction
• dynamic pages where ordinary fetches miss the real UI state

This capability is local-only. The browser bridge runs on the same machine as AI Now and is not exposed through Access Anywhere.

Read the full Browse Now guide for setup, limits, and CLI examples.

Plugins

Extend AI Now with connections to your other apps.

Custom MCP servers

Add a custom MCP server in AI Now → Plugins. Enter its HTTP MCP endpoint and required headers, test the connection, and enable it. For OAuth services, follow the sign-in and authorization prompts. Start with a read-only query to check the account and available tools.

The built-in runtime does not currently support STDIO MCP; do not enter a local launch command. For services such as Lemonade, chat model configuration and MCP tools are separate: configure the model in Settings and add its MCP endpoint separately for tools.

Built-in Plugins

Obsidian

1. Go to AI Now → Plugins
2. Enable Obsidian
3. Set your vault path

AI Now can now search and read your Obsidian notes alongside your memories.

Notion

1. Go to AI Now → Plugins
2. Enable Notion
3. Click Connect and authorize in the browser

AI Now can search and read authorized Notion content. Pages stay in Notion and do not automatically sync into the Library.

Apple Notes (macOS)

1. Go to AI Now → Plugins
2. Enable Apple Notes
3. Grant permission when prompted

Grant the system permissions shown in plugin settings. This plugin reads notes; it does not automatically import or sync them into the Library.

Ask about Mem, manage settings, and submit feedback

Use an ordinary AI Now task for these actions. You do not need a separate Ask Mem entry or an installed help skill.

Ask how to use Mem

How do I import Codex conversations into Mem? Use the current documentation to explain the first step and how to verify the import.

AI Now can consult product documentation and guide you through the steps. Account login, device permissions, and manual actions must still be completed when required. Finding instructions does not prove that a connection works.

Check or change settings

Check my interface and AI output languages. Change the AI output language to Chinese, keep other settings, and read back the saved value.

You can also ask about models, import configuration, or processing status. AI Now first checks the supported settings and permissions for the current connection, changes the requested fields, and verifies the result. Interface language, AI output language, task model, and background model are separate settings; specify which one you want to change.

Some saved settings need a restart before the running backend uses them. AI Now can schedule an explicitly requested restart only for a backend managed by the local desktop app. Remote and Cloud connections cannot restart your local backend. After reconnecting, check the effective state. Install or reinstall local models manually in Settings → Models.

Investigate a problem

Why did my AI Now task fail just now? Check available runtime state, configuration, and relevant diagnostics. Separate confirmed facts from possible causes and explain the next step.

AI Now can start with read-only checks of versions, runtime state, active tasks, resource usage, and relevant logs where the connection supports them. First identify whether it is checking the client or the Mem server. Process samples cover the Mem backend and its descendants, not the entire desktop app or WebView. A single memory reading does not prove a leak.

The result should state what was checked, what was found, what remains unknown, and the next step supported by evidence. Diagnosis alone does not authorize a restart, deletion, configuration change, or feedback submission. Explicitly request those actions when needed.

Submit product feedback

The task failed while generating a report just now. Please submit a bug report with this task's diagnostic information and return the feedback number.

AI Now explains what it will send before submission. Asking for help or diagnosis alone does not submit a report. Where supported, local and self-hosted connections can include App/OS versions, redacted logs, and images you select and attach to the current conversation. There is no automatic screenshot capture. Bug reports normally include recent redacted logs; feature requests and questions omit logs by default. You can request "no logs" or "no images." Cloud feedback uses its workspace's supported content and does not have the same local attachment capabilities.

A successful submission returns a feedback number and the attachments actually included. If the result is unknown or times out, check for an existing report before sending again. A received report does not mean the problem is fixed.

Session Management

Conversations are saved automatically. Click a previous session to resume, or create new sessions for parallel workstreams. Each session maintains its own history.

Recover from interruptions and failures

After an interruption, read the failure message and expand Details. Use Continue when offered, or edit and resend your message. Check files or memories already saved before repeating writes, sending messages, or submitting feedback. Stopping a task does not undo completed operations.

For support, use Copy diagnostics and review the content before sending it. You can also explicitly ask AI Now to submit feedback and return its number. Do not include passwords, API keys, or unrelated private content.

Auto-Approve Mode

Enable Auto to skip confirmation prompts for file operations and other actions.

Auto does not grant account permissions or enable unsupported backend operations. Scheduled tasks do not inherit chat auto-approval.

Using AI Now from Another Device

If your main Nowledge Mem runs on an always-on machine (a Mac Mini, a server, your office desktop), you can use AI Now from any other device that connects via Access Anywhere.

This is not a separate second knowledge base. You are using the same Mem from another client. If you want the broader model, read Sync Across Devices.

How it works:

• Connect from the desktop app, the web app at /app, or the mobile app through Access Anywhere
• AI Now runs on the Mem server you connected to, using that server's LLM configuration
• Your session list, message history, and pasted images stay with the server, so you can switch clients and continue the same work
• Your memories, threads, and Library are read from the same knowledge base through the secure connection

A web client connected to self-hosted Mem differs from Hosted Cloud. Hosted Cloud focuses on conversations with your knowledge and does not provide local files, Shell, Browse Now, or local app plugins. File analysis, web research, and scheduled tasks need backend support. Check the current connection if a control is absent.

Tips

• Be specific: "What did we decide about the database migration last month?" beats "database stuff"
• Attach context: drop files or mention saved content with @ for better results
• Use sessions: separate sessions for different projects or topics

Next Steps

• Access Anywhere: Use Mem from any device, the foundation for multi-device workflows
• Sync Across Devices: Understand the single-Mem, multi-client sync model
• Model providers: Check available providers and AI Now model selection guidance
• Connectors: Connect your AI tools and capture paths
• Background Intelligence: How your knowledge grows on its own