Strawberryforproductandengineeringteams

Think of Strawberry as a new hire for your engineering and product team. Capable and eager, but it needs your context first. Give it that and it takes on the repetitive half of the work: triage, monitoring, scraping, and research you never have time for. The more you tell it, the better it gets.

Setup is quick. Import from your previous browser and bring over your memories from ChatGPT, Claude, or Gemini. Connect the apps where your work lives: GitHub, Linear or JIRA, Slack, your email, and any Sheets, Notion, or databases that hold tickets - the MCP integration guide walks through connecting your tools. Then talk to it, by typing or by voice, and tell it what's on your plate this week.

Five concrete use cases

Build a data moat from sources behind a login

Turn a source that resists scraping into a clean dataset you own.

• Set the scope: which source, which fields, how many rows, and where it lands (a Sheet, Notion, CSV, or your own database). Flag quirks like pagination, infinite scroll, or login walls.
• Your companion scrolls and paginates the source, handles JavaScript-heavy pages, extracts the rows, and can run parallel agents to enrich each one. It shows you a clean table with a fit check before anything exports.
• Save it as a skill so you never re-explain the source, then set a routine that re-scrapes on a schedule and flags what changed. Data you capture again and again becomes a moat.

Triage a bug, and only file it if it's new

Stop filing duplicates. Your agent checks before it creates.

• Tell it where bugs live, the team, project, labels, severity scale, and the template you use. Connect Linear or JIRA.
• Describe the bug or paste the repro. Your agent searches existing issues for a duplicate and links you to it if one exists. If it's new, it drafts a clean ticket with steps, severity, and labels for you to approve. It suggests, you confirm.
• Save it as a skill, then set a signal-based routine that triages reports as they land and pings you only when something new and serious shows up.

Monitor your repos and CI

Get a single read on what needs attention instead of opening twenty tabs.

• Point it at the repos that matter and what you care about: open PRs, stale reviews, new issues, releases, build status, coverage gaps. Connect GitHub.
• Your agent pulls the state across repos and gives you a digest of what needs review, what's blocking, what shipped, and which builds are flaky. Ask follow-ups on any PR, or have it draft release notes from the PRs since the last tag.
• Save it as a skill, then set a routine that posts a daily digest to Slack or chat.

Track what competitors are shipping

Know what changed across your market without reading every changelog yourself.

• Name the competitors and the signals: changelogs, pricing, features, positioning, leadership changes, funding or M&A. Tell it where to save findings.
• Your companion reads changelogs, pricing pages, docs, blogs, and news, then presents a structured comparison and flags what changed since last time.
• Save it as a skill, then set a routine that re-checks weekly and surfaces only the diffs worth your time.

Hear what people are saying about your product

Turn scattered mentions into themes you can act on, with early warning when something spikes.

• Name your product and the sources to watch: Reddit, X, Hacker News, review sites, app stores, support tickets, NPS or CSAT. Set the themes and sentiment buckets you care about.
• Your agent pulls recent mentions, tags sentiment the way customers phrase it, clusters themes, and surfaces the loudest pain and praise. It flags a jump in a complaint before it becomes a fire.
• Save it as a skill, then set a routine that summarizes new mentions on a schedule and alerts you on anomalies.

More ideas

A few less-obvious ways product and eng teams put their companion to work:

• QA before a launch. Crawl every page for typos, broken links, and inconsistent copy, and run a happy-path click-through each release. Stay in the loop on flaky flows like sign-in or checkout.
• Reproduce a reported bug. Walk a customer report end to end in the browser, capture screenshots, and attach it to the ticket.
• Skip the blank page. Draft a PRD or user stories from raw feedback, then audit your docs for dead links and outdated screenshots.
• Mine interview transcripts. Synthesize user interviews for patterns you missed, then draft a note to customers when their feature ships.
• Tidy the backlog. Turn a messy issue list into a prioritized, de-duplicated one.

Use Strawberry with your team

These playbooks get better when the whole team shares them. Share your companion, context, and skills so everyone inherits the same scraping, triage, monitoring, and research flows. A competitor diff or a bug triage one teammate runs is available to the rest of the team.

See the Strawberry for teams page for full setup.

Related

• Strawberry for teams
• Get started with Strawberry
• Strawberry for meetings