Browser automation · Checked 6 August 2026

Strawberry Browser: Give an AI a repeatable browser job

Research, extract, and work across sites from a desktop browser built around AI tasks.

Strawberry Browser is a desktop browser for repeatable AI research, extraction, prospecting, and connected-app workflows on macOS and Windows.

Type
desktop app
Works on
macOS, Windows

Open Strawberry Browser ↗︎

Use it for a job with a finish line

Strawberry is a desktop browser with AI companions that can read pages, extract structured information, research across sites, and work with connected apps. The useful distinction is the task shape. It is aimed at jobs such as building a competitor brief, finding a defined set of prospects, extracting every plan from a pricing page, or summarising a fixed group of sources.

Give it a small output contract on the first run. For a competitor check that could be five companies, their current pricing URL, one quoted product claim, and the date each source was checked. A defined table is easier to inspect than “research this market” and exposes missing sources quickly.

The company publishes a 12-task benchmark specification covering research, prospecting, CRM work, web extraction, and other browser jobs. The specification asks testers to record blockers and source quality as well as completion. Those are sensible checks even if you never use its published scores.

The free plan is enough for a controlled trial

The free plan included 2,000 credits per month when checked on 6 August 2026. Ordinary browsing did not consume credits; companion chat and companion browsing did. Paid plans started at $20 per month for 8,000 credits, with larger plans at $100 and $250 per month.

Version 0.1.22 was available for macOS and Windows. We completed the Mac download flow, verified the disk image checksum, and confirmed that the app carried an accepted notarized Developer ID signature. We did not launch the app, create an account, or complete one of its research tasks, so this page does not claim a successful end-to-end automation.

Local passwords do not make every AI task local

Strawberry says passwords, cookies, and some operational data stay on the device. Its privacy policy gives a fuller picture. User content can also be stored in cloud infrastructure and sent to third-party AI providers while an agent completes a task. The policy says model training on user content is off by default, but its measures for keeping credentials and sensitive data out of AI requests are not absolute.

An optional Activity Memory feature can capture periodic screenshots from the screen, including other apps. It is disabled by default and requires an opt-in, with pause, export, and deletion controls. Leave it off during an initial test.

Start on public web pages with no personal data. Review the extracted table and every important source before letting Strawberry touch a CRM, inbox, private document, or logged-in prospecting service.

Sources

Swipe the best AI skills, prompts, and workflows.

The newsletter to learn AI by stealing the cool ideas, skills, and tools you didn't know you needed.

One more step.

Check your inbox to confirm your subscription. Can't find it? Check your spam or junk folder.