Guide

Email Analytics with Manus AI

Use Manus AI to analyze your email patterns — volume by sender, response times, busiest hours, and unread trends. Nylas CLI pulls the data, Manus generates the report. Works across all major email providers.

Written by Hazik Director of Product Management

Reviewed by Nick Barraclough

Verified — CLI 3.1.1 · Gmail, Outlook · last tested April 11, 2026

Why email analytics

Email analytics turn raw inbox data into actionable patterns — top senders by volume, peak delivery hours, unread age distribution, and response-time estimates. The average knowledge worker receives 120 emails per day according to the Radicati Group, yet most email clients show only a chronological list with no aggregation tools built in.

Manus AI can pull inbox data through the Nylas CLI, analyze it in its sandbox, and generate structured reports — tables, ranked lists, or downloadable spreadsheets — without you writing any code. The CLI returns JSON that Manus parses directly, so there is no manual export step between your mailbox and the analysis.

Prerequisites

Setting up email analytics with Manus AI requires three components: a Manus account, the Nylas CLI Skill installed in that account, and at least one connected mailbox. The entire setup takes under 5 minutes if you already have a Nylas grant configured.

  • A Manus account — sign up at manus.im. Free and paid plans both work for analytics prompts.
  • Nylas CLI Skill installed in Manus — follow the Manus AI Skills guide to set up the Skill that gives Manus email and calendar access.
  • A connected mailbox — at least one grant (Gmail, Outlook, Exchange, Yahoo, iCloud, or IMAP) configured in your Nylas dashboard.

Pull inbox data

The Nylas CLI's nylas email list command is the data source for every analytics prompt. It returns structured JSON containing sender address, timestamp, subject line, read/unread status, folder, and thread ID — the fields Manus needs to compute volume breakdowns and timing patterns.

Each JSON object in the response maps to a single email message. The --limit flag controls how many messages the CLI fetches in one call; 100 messages is enough for a weekly snapshot, while 500 covers roughly a month for most inboxes. The command below fetches 100 messages as JSON.

nylas email list --limit 100 --json

Manus runs this command inside its sandbox and receives the full JSON array. From there, it can slice and group the data any way you ask. Larger datasets use more Manus credits, so start with 100 and increase only when you need a wider time window.

Volume by sender

Volume-by-sender analysis ranks every email address that wrote to you by message count over a given period. Research from Sanebox found that 62% of email volume is "not important" — newsletters, automated alerts, and marketing blasts. Identifying those high-volume senders is the first step toward filtering or unsubscribing.

The prompt below asks Manus to fetch your recent messages, group them by the from field, and return a ranked top-10 table. Manus calls nylas email list internally, so you don't need to run the CLI command yourself.

Prompt: "Show me who emails me the most this month. Rank the top 10 senders by message count."

The resulting table typically reveals that 3-5 senders account for the majority of your inbox. Automated notifications and mailing lists tend to cluster at the top, while individual colleagues appear further down with fewer but higher-priority messages.

Busiest days and hours

Timing analysis breaks your email flow into day-of-week and hour-of-day buckets so you can align focus blocks with low-volume windows. A 2023 Superhuman study found that Tuesday 10 AM is the single highest-volume hour for business email globally, but individual patterns vary widely depending on industry and time zone.

Manus extracts the date timestamp from each message in the Nylas CLI output, buckets it by weekday and hour, and returns a heatmap or table showing your personal peaks. The prompt below requests both dimensions in one pass.

Prompt: "When do I receive the most emails? Break it down by day of the week and hour of the day."

The output often reveals that a single weekday accounts for 25-30% of total weekly volume. Use that information to schedule deep work on quieter days and batch-process email during peak windows.

Unread message analysis measures inbox pressure — how many messages are waiting, how long they've been waiting, and which senders contribute the most unread volume. A Microsoft study reported that the average Outlook user has over 100 unread emails at any given time, so tracking unread age helps separate genuine backlog from low-priority noise.

The Nylas CLI's --unread flag filters the response to only unread messages. Combined with --json, this gives Manus a dataset that includes the timestamp of each unread message, which it uses to calculate age and group by sender. The command below fetches all unread messages as JSON.

nylas email list --unread --json

Once Manus has the unread dataset, you can ask it to group by sender and sort by message age. The prompt below returns a sender-grouped breakdown with the oldest unread timestamp for each sender, so you can spot messages that have been sitting for days or weeks.

Prompt: "How many unread emails do I have, and how old are the oldest ones? Group by sender."

The results surface messages you've been avoiding and help you decide what to read, archive, or delegate.

Generate a report

Manus AI can combine multiple analytics passes — volume, timing, unread status, and response time — into a single downloadable report. A typical weekly report covers 5 dimensions: total volume (sent and received), top senders, peak hours, unread backlog, and estimated reply latency. Generating all five from one prompt saves Manus credits compared to running each analysis separately.

The prompt below asks Manus to run multiple nylas email list commands with different filters, compile the results, and format the output as a spreadsheet. Manus supports CSV, HTML table, and Markdown output formats — specify the format you want in the prompt.

Prompt: "Create a weekly email report with:
- Total volume (received and sent)
- Top 10 senders by count
- Busiest day and hour
- Unread count and oldest unread
- Response time estimates for my replies

Format it as a downloadable spreadsheet."

Manus generates the spreadsheet as a downloadable file inside its sandbox. You can open it directly or save it for comparison with future weeks to track trends over time.

Contact analytics

Contact analytics cross-reference your email senders with your address book to find gaps — frequent correspondents who aren't in your contacts, CRM records missing recent email addresses, or unknown senders who deserve a contact entry. Salesforce research shows that sales teams lose 27% of contact data annually to job changes and turnover, so regular cross-checks keep your records current.

The Nylas CLI's nylas contacts list command returns your full address book as JSON, including name, email address, company, and phone number fields. Manus can then join this dataset with the email sender list from nylas email list to identify matches and misses. The command below fetches your contacts.

nylas contacts list --json

Once Manus has both the email and contacts datasets, ask it to match senders against contacts by email address. The prompt below requests a match report with suggestions for which unknown senders to add to your address book.

Prompt: "Match my top email senders to my contacts list. Show me who is in my contacts and who isn't. For unknown senders, suggest which ones I should add."

The match report highlights senders you email frequently but haven't added as contacts, which is useful for keeping your CRM or address book up to date.

Tips for better analytics

A few prompt-writing habits make Manus analytics more accurate and more credit-efficient. Specifying exact time ranges, requesting structured output formats, and batching multiple questions into a single prompt all reduce the number of CLI calls Manus needs to make. Each nylas email list call counts against your Manus credits, so fewer calls means lower cost per report.

  • Be specific with time ranges — "this week" and "last 30 days" give Manus clear boundaries for filtering.
  • Request specific formats — ask for "a table", "a CSV", or "a bullet list" to control how Manus presents results.
  • Combine prompts — batch multiple analytics questions into one prompt to reduce credit usage and get a unified report.
  • Save as a Skill — if you run the same analytics weekly, package your prompts as a Manus Skill so you can trigger them with a slash command.

FAQ

How accurate are Manus AI email analytics?

Manus works with the raw data returned by the Nylas CLI. Counts and timestamps are exact. Derived metrics like estimated response times depend on how well Manus can match replies to originals, which is reliable for direct threads but less precise for forwarded or merged conversations.

Does Manus store or share my email data?

Manus processes data inside an isolated sandbox that is destroyed after each session. Email content is not persisted or used for training. Your Nylas API credentials remain in the sandbox and are not accessible outside of it.

Can I schedule recurring email analytics reports?

Manus does not have built-in scheduling. However, you can save your analytics prompts as a Manus Skill and re-run them on demand. For automated recurring reports, pair the Nylas CLI with a cron job or CI/CD pipeline instead.

Next steps