Breaking

Rebrowse, an Open-Source Recorder That Turns Browser Actions Into Agentic Workflows, Debuts on GitHub

July 16, 2026 at 13:29 EDT

AI Agents
Open Source
Software Dev & Coding

A new open-source tool called Rebrowse aims to let anyone build automated browser workflows simply by recording their clicks and interactions, converting those recordings into deterministic agentic flows that can be replayed in the cloud with a single tap.

Open Source · Browser Automation

Record Your Clicks, Replay as an Agent

Rebrowse is a self-learning browser recorder that turns a captured session into a deterministic, repeatable workflow that runs in a cloud browser — no scripting or node-wiring required.

~20×

faster execution (claimed)

95%

accuracy / success rate

~1 min

to create a workflow

127

GitHub stars at sharing

Deterministic vs. LLM-per-run

Speed advantage claimed for replaying a recorded flow instead of re-reasoning every step. Each block ≈ 4×.

1×

Typical AI copilot run

~20×

Rebrowse replay

How it works · "Loom for workflow automation"

Record

Chrome extension captures clicks & interactions, plus optional voice intent

→

Generate

Session becomes a deterministic agentic flow via the OpenAI API

→

Replay

Runs in a cloud browser via Playwright — with parallel runs in v0.2.0

Where it shines

Repetitive tasks for QA, marketing & ops
Fast prototyping and easy sharing
Stable sites where the UI rarely changes
Intuitive for non-developers

Where it breaks

Fragile when element IDs or layout shift
Poor fit for OAuth / multi-factor logins
Weak on branching, logic & data transforms
Performance claims not independently verified

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.