Browser4 Agentic Browser
A browser automation CLI for interactive workflows, web crawling, and structured page extraction.
- Source repo
- platonai/Browser4
- Stars
- ★ 1.2k
- Last updated
- 2d ago
- License
- Apache-2.0
- Primary language
- Kotlin
- FA score
- 65/100 · Some gaps
At a glance
- How it runs
- Works with
- Universal · cross-platformOpenAI API
- Cost
- Free software; you pay for model usage
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- A QA engineer validating a logged-in form flow can inspect snapshot references, fill fields, click controls, and wait for a success message or network-idle state.
- Not a fit if
- Users unwilling to install JDK, Chrome, and Rust build prerequisites
- Anyone expecting built-in bypass of CAPTCHAs or anti-bot defenses
- Website
- browser4.io ↗ set on GitHub
- Maintenance
- Commits in the last 30 dayslast commit 2026-11-09latest release v4.16.4 (2026-11-08)+0.6 ★/day over 7 days
- Source review
- 65/100 · Some gaps
What does this agent do, and when should you use it?
Browser4 is a browser automation and web-data processing project centered on the browser4-cli command-line interface. Its documented architecture combines a Rust CLI, the Kotlin/Spring browser4-rest backend, a Chrome DevTools Protocol browser driver, and browser4-agentic tooling connected through MCP over HTTP. It supports live page interaction through accessibility-tree element references, alongside stored HTML snapshots for CSS-selector and X-SQL extraction. Crawl, swarm, loop, and batch commands cover larger or recurring jobs, while screenshots, PDFs, HTML exports, and browser-state operations provide outputs and persistence. The CLI can be installed with npm, or the project can be built from source into a unified Browser4.jar launcher.
browser4-cli open or goto opens or reconnects a browser session. snapshot exposes interactive page elements as references such as e15, which commands including click, fill, type, press, select, and wait can use to operate the page. For extraction, htmlsnapshot capture stores the DOM; htmlsnapshot get, htmlsnapshot inspect, and htmlsnapshot query --sql @query.sql read text, attributes, or correlated records from that snapshot, while extract and summarize require an LLM provider key. crawl processes a URL or seed file, and swarm create enables parallel submission or X-SQL query work that is retrieved with status and result commands. The CLI also creates screenshots, PDFs, and exported HTML, persists cookies and localStorage, and calls the backend through MCP over HTTP.
- A QA engineer validating a logged-in form flow can inspect snapshot references, fill fields, click controls, and wait for a success message or network-idle state.
- A data analyst extracting title, price, and URL from each search-result card can capture an HTML snapshot and query correlated fields with X-SQL.
- A scraping engineer with a file of known URLs can create a swarm and run the same SQL extraction across pages in parallel.
- A researcher collecting pages reachable from a starting site can configure crawl with a depth and an outgoing-link selector.
- An operations user monitoring a page or health endpoint on a schedule can use loop to repeat a browser task or shell command.
Key setup facts
Pulled from the install and usage notes below. The project README has the current values.
- Install / run
npm install -g browser4-cli
- Environment variables
OPENAI_API_KEYOPENAI_BASE_URL- Runtime
- Java 17+
How do you install or deploy this agent?
Global CLI installation requires Node.js:
npm install -g browser4-cli
browser4-cli installVerify the first session with:
browser4-cli open --headed https://browser4.io
browser4-cli snapshot --boxesAI commands such as extract, summarize, chat, and agent run require a supported provider key. For an OpenAI-compatible provider, configure OPENAI_API_KEY, OPENAI_MODEL_NAME, and OPENAI_BASE_URL. Source builds require Git, JDK 17+ (21+ recommended), and Chrome or Chromium; building the CLI from source also requires a Rust toolchain.
How do you use this agent?
For an interactive flow:
browser4-cli goto https://example.com/login
browser4-cli snapshot -i --boxes
browser4-cli fill e3 "[email protected]"
browser4-cli fill e4 "secret" --submit
browser4-cli wait --load networkidleFor structured extraction, run browser4-cli htmlsnapshot and then browser4-cli htmlsnapshot query --sql @query.sql. Process a URL list with browser4-cli crawl --seed-file urls.txt --depth 0 --sql @query.sql, or create a swarm first when parallel processing is needed.
What are this agent's strengths and limitations?
- It deliberately separates accessibility-tree references for interaction from stored HTML snapshots and X-SQL for repeatable DOM extraction after dynamic page work.
- The Rust CLI, Kotlin/Spring backend, and CDP driver have documented boundaries and communicate through MCP over HTTP.
- crawl, swarm, batch, loop, and named sessions span single-page interaction, bulk crawling, parallel jobs, and recurring execution.
- It includes cookie and storage controls plus HTML, screenshot, and PDF outputs for stateful automation workflows.
- AI-powered extract, summarize, chat, agent run, and X-SQL llm_* functions depend on external LLM provider credentials.
- Source builds have several prerequisites: JDK, Chrome/Chromium, PowerShell 7 in the documented Linux/macOS prerequisites, and Rust for the CLI build.
- The CLI persists session state, cookies, and localStorage, so adopters need their own approach to credential and local-state handling.
- The documented command set does not establish a general guarantee for anti-bot defenses, CAPTCHAs, or site-specific authorization failures.
How does this agent compare with similar options?
WebMiner is presented as a complementary path: Browser4 can acquire individual or bulk HTML pages, while WebMiner clusters downloaded HTML locally to create interactive reports and Excel spreadsheets without LLM-token use. Browser4 covers browser interaction and acquisition; WebMiner emphasizes offline HTML batch processing.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Browser4 Agentic Browser This agent | 65 · Some gaps | CLIFree + model costs | ★ 1.2k | 2d ago | Kotlin | OpenAI API |
| Obscura Headless Browser | 68 · Some gaps | CLIFree | ★ 32k | 1d ago | Rust | Claude.ai |
| Opensteer | 54 · Major gaps | CLIFree | ★ 211 | today | TypeScript | — |
| Open Browser Use | 65 · Some gaps | CLIFree | ★ 419 | 8d ago | JavaScript | Codex · Claude Code |
How does FollowAgents rate this agent?
Why each dimension lost points
The CLI, MCP, browser driver, and coding toolchain expose broad capabilities, while the packet does not show task-specific permission controls, so least privilege is only partly demonstrated. Uninstall summarizes deletion scope and asks for confirmation, providing limited confirmation safeguards. The docs recommend environment variables for API keys, describe a commented-by-default config template, and report the selected provider; however, a log example includes an apiKey field and the packet does not explain redaction, so sensitive-data handling earns only a low score. Installer docs identify download sources and options, but the packet does not show checksums, signatures, or dependency vulnerability management. Visible side effects include runtime downloads, attempted Chrome installation, user PATH changes, and config-file creation; some can be skipped or dry-run, but confirmation and recovery coverage is incomplete. The uninstall prompt and versioned runtime directories provide some recovery guidance. The Apache license is clear. Publisher identity is unknown per the prompt, and repository evidence does not fully establish attribution or maintenance responsibility.
The README, skills, and install docs give largely consistent coverage of commands, operating modes, environment requirements, and recovery. Some configuration material appears interleaved and portions are truncated, so consistency across the full product cannot be confirmed. The product depends on Chrome/Chromium, Java 25, Maven, Rust, and external download sources; prerequisites and installation paths are documented, but availability and a complete offline path are not demonstrated. Concrete symptoms, diagnostic commands, and recovery steps support a good static score for failure messages.
The packet distinguishes visible browser use for humans from headless-by-default agent use, and covers interaction, extraction, crawling, batching, MCP, and development scenarios. Skill descriptions specify triggers and allowed tools, making intended use relatively precise. Docs identify boundaries such as LLM-key requirements, same-origin iframe support, remote upload paths resolving on the backend host, and network-idle not guaranteeing rendered results, though the list is not exhaustive. Installation and development guidance covers Windows, Linux, macOS, Git Bash, CI, and containers.
The README provides a contents list, tool-selection guide, command reference, architecture, and module map; skills are organized into a decision layer and topic references. Install, upgrade, uninstall, and cross-platform installer guidance is detailed, with many copyable examples and troubleshooting steps. Command and module naming is generally recognizable and stable, but the packet does not provide enough complete release history for full marks. Documented limitations include browser-version compatibility, network-idle semantics, startup latency, and platform differences, though the coverage is not exhaustive. The complete license text matches the metadata. CI, scheduled maintenance, and version-consistency checks are shown, but the packet does not name a clear maintenance owner or support commitment; unverified publisher identity is not itself a deduction.
The CLI offers readable snapshots, structured JSON, file export, diagnostic status, and asynchronous task identifiers, but the packet does not establish that every output integrates cleanly downstream, supporting a moderate score. Browser interaction and extraction, zero-LLM X-SQL/HTML pipelines, batch and swarm processing, and experience memory provide clear automation value. Runtime downloads are about 200 MB, source startup can take minutes, and some capabilities require an external LLM or browser; costs are disclosed but benefits and resource tradeoffs are not quantified.
Docs frequently point to specific config files, commands, endpoints, modules, and failure cases, making many claims traceable; the README, skills, installer docs, and CI corroborate several workflows. The packet describes mechanisms and checks, but this review did not execute code, and CI configuration or documentation does not prove this revision passed a build or tests. Claims are often accompanied by mechanism descriptions, but promotional throughput figures lack independent support in the supplied sample, and fact versus inference is not consistently separated.
- Browser4 can control real browsers, fill forms, upload local files, visit sites, and run large crawls; review the target site, session, file paths, and external effects before use.
- The runtime and installer scripts download components from the network; the packet does not show signature or checksum verification. Review download sources before deployment and handle API keys and potentially sensitive logs carefully.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
v4.16.4 🚀 Browser4 4.16.4
Release notes excerpt
🚀 Browser4 v4.16.4 Release Notes Release Date: 2026-11-08 15:00:33 UTC Built by: @galaxyeye > ### ☕ Java requirement — depends on how you install > > | Install mode | Java needed? | > |--------------|--------------| > | Self-contained bundle — browser4-cli install | None — the bundle ships its own jlink-built JRE | > | browser4-cli binaries / npm install -g browser4-cli | None — native binary | > | Standalone Browser4.jar (java -jar) | Java 25+ | > | WebMiner (browser4-cli webminer) | JDK 17+ — auto-detected (the bundle's JRE also satisfies it) | …View on GitHub ↗ -
v4.16.3 🚀 Browser4 4.16.3
Release notes excerpt
🚀 Browser4 v4.16.3 Release Notes Release Date: 2026-11-07 06:05:15 UTC Built by: @galaxyeye > ### ☕ Java requirement — depends on how you install > > | Install mode | Java needed? | > |--------------|--------------| > | Self-contained bundle — browser4-cli install | None — the bundle ships its own jlink-built JRE | > | browser4-cli binaries / npm install -g browser4-cli | None — native binary | > | Standalone Browser4.jar (java -jar) | Java 25+ | > | WebMiner (browser4-cli webminer) | JDK 17+ — auto-detected (the bundle's JRE also satisfies it) | …View on GitHub ↗ -
v4.16.2 🚀 Browser4 4.16.2
Release notes excerpt
🚀 Browser4 v4.16.2 Release Notes Release Date: 2026-11-06 13:23:54 UTC Built by: @galaxyeye > ### ☕ Java requirement — depends on how you install > > | Install mode | Java needed? | > |--------------|--------------| > | Self-contained bundle — browser4-cli install | None — the bundle ships its own jlink-built JRE | > | browser4-cli binaries / npm install -g browser4-cli | None — native binary | > | Standalone Browser4.jar (java -jar) | Java 25+ | > | WebMiner (browser4-cli webminer) | JDK 17+ — auto-detected (the bundle's JRE also satisfies it) | …View on GitHub ↗
FAQ
Which features require an LLM key?
Can it work with authenticated websites?
How should I process many URLs?
Is there a no-token route for structuring downloaded pages?
Is Browser4 Agentic Browser free?
Is Browser4 Agentic Browser open source?
Is Browser4 Agentic Browser safe to use?
Is Browser4 Agentic Browser still maintained?
Who makes Browser4 Agentic Browser?
Thanks — we use this to improve the page.