Strawberry is a Chromium-and-Electron browser from Dendrite Systems built around AI Companions — agents that research prospects, update spreadsheets, and replay recorded workflows while you watch. We spent hands-on time with its BETA release to test how well the self-driving promise holds up, and where a free alternative like Tabbit fits.
Editorial Verdict
Promising agentic concept / early-beta polish gap
Strawberry deserves credit for making browser agents visible, pausable, and approval-gated. Companions handle research, lead sourcing, and recurring routines inside your logged-in tabs, and the no-code workflow recording is clever. But heavier tasks lag, complex instructions still wobble in this BETA, and the credit meter ($10 per 1,000 credits after 500 free) quietly taxes the heaviest users. For teams that want the same approval-based automation without the meter or the framework tax, Tabbit is the stronger daily driver.
Jump to section
Product Identity
Strawberry positions itself as “the self-driving browser”: a standalone desktop browser where AI Companions click, fill forms, extract data, and manage tabs for you. It is developed by Dendrite Systems, runs on Electron + Chromium, and currently ships as a BETA for macOS and Windows.
Strawberry is built by Dendrite Systems, a Swedish team that has run three Product Hunt launches, most recently positioning the browser as the place “where teams and agents work together.”
Under the hood, Strawberry is an Electron app rendering with Chromium. That enables deep desktop integration, but reviewers consistently report more lag and memory use than native Chromium browsers like Chrome, Edge, or Tabbit.
There is no official Linux or mobile build — a gap users actively search for. Per the company FAQ, passwords, chats, and history stay local on your device.
Hands-On Findings
We ran Strawberry’s headline workflows side by side with Tabbit’s Agent Mode. Three patterns stood out.
Give a Companion a plain-English brief — “source 50 leads and add them to our CRM” — and it reads playbooks, searches LinkedIn and Crunchbase, scores companies against your ICP, and writes rows into Google Sheets. Impressive while the run stays simple; heavier sweeps get slow.
Strawberry’s routines trigger on events — “whenever a support email arrives, run the resolution playbook” — and bring results back for review. It is the closest thing to an always-on digital coworker we have tested, provided you budget credits for every activation.
Strawberry pauses before sensitive actions like sending email. Tabbit implements the same safety idea with a live, step-by-step execution plan — Load Skill, Update Plan, Click, Snapshot — plus free Agent Mode during public beta, so approvals never come with a meter attached.
Evaluation Matrix
Our scoring methodology tests the dimensions power users care about when an AI browser becomes their primary workspace.
| Feature / Dimension | Testing Methodology | Strawberry Performance | Tabbit Advantage |
|---|---|---|---|
| Agent transparency | Watching, pausing, and replaying agent runs; inspecting what the AI is about to do | Excellent: visible companion steps, pause and replay praised across user reviews | Live execution plan with per-step snapshots and explicit approval cards |
| Browsing performance | Page load, scroll smoothness, and memory footprint during long sessions | Electron overhead: noticeable lag and higher resource use versus native browsers | Native Chromium engine tuned for speed with lower memory overhead |
| Workflow reliability | Multi-step runs on complex sites with forms, filters, and dynamic DOMs | Solid for simple briefs; complex instructions can miss outcomes in this BETA | Deterministic Skills library plus agent improvisation, tested on real workflows |
| Model choice | Which AI models power the agent and the assistant chat | Tied to Strawberry’s own model routing; no user-facing model picker | Route freely across Claude, GPT, Gemini, DeepSeek, Kimi, Qwen and more |
| Cost model | Real monthly cost for daily research, CRM updates, and outreach drafts | 500 free credits, then $10 per 1,000; heavy teams land on $85–$250 plans | Public beta: Agent Mode, deep research, and multi-model chat free |
| Platform coverage | macOS, Windows, Linux, and mobile availability | macOS and Windows desktop only; no Linux or mobile builds | Native Windows and macOS builds with full web compatibility |
Pros & Cons
Pricing Reality Check
Browsing itself is free — credits are consumed only when Companions act. The meter is the catch: a single research-and-CRM update run can burn hundreds of credits, and top-ups cost $10 per 1,000.
| Plan | Price | Monthly Credits | Best For |
|---|---|---|---|
| Free | $0 | 500 | Trying companions on light tasks |
| Intern | $20/mo | 6,000 | Solo users automating daily chores |
| Part-time | $85/mo | 25,000 | Delegating 15–20 hours of work weekly |
| Full-time | $250/mo | 75,000 | Automating 40+ hours of complex workflows |
| Team | $35–$350/seat | 10,000–100,000 | Shared pools with admin controls |
All Strawberry prices from strawberrybrowser.com/pricing (October 2026); $10 per extra 1,000 credits on every plan.
Target Audience
If your work is dominated by event-triggered loops — inbox triage, CRM hygiene, candidate sourcing — Strawberry’s routines and shared companions are genuinely well designed. Budget for the Part-time tier or above.
Great fit if credits fit the budget
If the browser is your primary workspace, Electron lag compounds. Tabbit gives you the same approval-based agent automation on a fast native Chromium base that stays responsive with dozens of tabs open.
Tabbit wins on daily-driver speed
Strawberry routes through its own stack with no picker. Tabbit lets you send the same task to Claude, GPT, Gemini, DeepSeek, or Kimi and compare answers side by side — free during beta.
Tabbit wins on model freedom
The Alternative
Tabbit is an AI-native browser built on a fast Chromium core for both Windows and macOS. Its Agent Mode plans every action before it runs, shows a live step-by-step execution log, and asks for approval before anything sensitive — the same safety philosophy as Strawberry, without credit counting.
Add multi-model routing, one-click migration from Chrome, and a Skills library of reusable browser automations, and Tabbit becomes the natural next stop for anyone outgrowing Strawberry’s BETA.
Frequently Asked Questions
Download Tabbit and put approval-based AI agents to work on your real browser tasks — free while we’re in public beta, on macOS and Windows.
Free during public beta · macOS & Windows · No credit card required