Getting started with Span: for engineers

Last updated: November 2, 2026

Span turns data from tools you already use, such as your Git host, issue tracker, and AI coding tools, into a picture of how your work flows: where your time goes, what slows you down, and what you've shipped.

Why it matters for you

See where your pull requests wait. PR lifecycle splits your PR cycle time into Coding, Awaiting First Review, Reworking, and Idling, so you can tell writing time from waiting time.
Protect your focus time. If your organization connects calendars, you see your weekly focus time, fragmented time, and meeting time.
Get more out of AI coding tools. If your organization connects them, your metrics include your own AI usage, such as AI Days / week.
Make your work visible. Activity Log lists your commits, pull requests, reviews, and completed issues. Your Brag Sheet summarizes your work for a 1:1 or review once you have 3 months of history. Subscribe on your profile for a Weekly Update email.
Ask questions in plain English. If your organization has given you the Use Src permission, open Src in the sidebar and ask "Where did my pull requests wait last month?" Src cites its sources and runs with your Span permissions. With the default permissions, its answers cover the teams you're on and your own individual data, not your teammates'. Questions about repositories and code cover the whole organization. Your admin may also allow /src in Slack.

Where to start

Sign in. Span signs you in without a password. On the Log in to Span page, choose Continue with email or Continue with Google. With single sign-on (SSO), Span sends you to your identity provider; otherwise it emails you a 6-digit code. See Signing in to Span.
Confirm your pull requests show up. Open Activity Log on your profile and look for your recent pull requests. If they're missing, ask your Span admin to link your GitHub or GitLab username to your profile. See Linking people to their GitHub and GitLab accounts.
Look at your team's reports. Open your team from Organization → Teams, including the Targets your team has agreed to.
Open your Brag Sheet before your next 1:1. See Brag Sheets: summarize your work for reviews and 1:1s.
Take surveys when you're asked. Your organization may run DevEx or Pulse surveys through Span. Your answers show leaders where tools and processes slow the team down.

Getting around

Click your name at the bottom of the sidebar and choose My profile. Overview summarizes your week; Metrics, Activity Log, and Brag Sheet go deeper. The sidebar groups the rest under Organization, AI Effectiveness, and Insights; Organization → Teams has your team.

What you see depends on your organization's settings. When a report says "You are seeing a limited set of data based on your permissions.", click Learn more to open My Permissions.

If your organization gives you access to Traces

Engineers don't see agent traces by default. With access, you can review your own sessions and their evaluation scores:

Compare your sessions. Look at sessions that went well next to ones where you kept correcting the agent.
Write down what you keep re-teaching. If you type the same instruction session after session, add it to your repository's CLAUDE.md or AGENTS.md, or to a skill, and open a PR.

Your data and your privacy

Span focuses on team-level patterns. Your organization decides who sees individual data. With the default roles:

You see your own individual data and team-level reports for the teams you're on.
Your manager sees individual data for the people on their teams and in their reporting line, plus team-level reports for all teams.
Your organization's admins manage roles and permissions.
Some team views list pull requests by author. For example, a team's Targets page lists pull requests that missed a target.

Your view may differ depending on your organization's settings. Check My Permissions to see your access.

What Span reads: pull requests, commits, reviews, comments, and diffs from your Git host; issues and their history; and, if connected, AI tool usage and spend and calendar event times and attendees. Span redacts Google Calendar event titles and descriptions before storing them; from a description it keeps only links to the recruiting tools Greenhouse, Ashby, and Lever. For Microsoft Outlook, it keeps event subjects and redacts event bodies.

If your organization collects agent traces

Span records sessions in supported AI coding tools through auspex, Span's agent. auspex adds no Git hooks and writes nothing into your repositories.

Your prompts are yours. By default only you can read your prompts and transcripts (your organization can change this).
Others may see facts about your sessions (such as cost, model, and linked pull requests) if they have permission to view traces.
Redaction: secrets and personal data, such as keys and email addresses, are redacted before upload. It's a strong filter, not a guarantee.
Pausing collection: your organization can tell you whether you can pause collection, and how.
Trace Screening, if your organization uses it: personal or sensitive sessions are restricted so only you can see their contents, regardless of other people's permissions. Span doesn't expose those contents to anyone or anything other than you, including analysis such as evaluations, coaching, and Harness health. Others still see that the session exists, who authored it, and its model, cost, and duration. See Trace Screening: Automatically restrict personal or sensitive traces.

Surveys: in a Transparent survey, the people running it see who answered what; an Anonymous survey shows no results until at least 4 people respond.

Things to know

Data isn't live. New work shows up within a few hours.
History starts at the start date set for your organization.
Missing work? Your GitHub or GitLab username must be linked to your profile, and only people marked as Contributors count in per-person "/ week" metrics. Ask your admin to check both. See What determines contributor status in Span user profiles?
Out-of-office days leave "/ week" metrics. A weekday counts as out of office when calendar out-of-office events cover more than 4 hours between 8am and 6pm local time, or, if your organization uses an HRIS for time off, when it records an absence. Those days leave the denominator, and pull requests and commits from calendar out-of-office days leave the count. For example, if you merge 6 pull requests in a week with 2 out-of-office weekdays, PRs merged / week is 6 ÷ 5 active days × 7 = 8.4, not 6.
Look at trends, not single weeks. A week with a holiday, an incident, or one large migration PR can look very different from the rest.

Need help?

Ask your internal Span admin first: they manage access, permissions, and account linking. Otherwise, reach out to support@span.app.