Capture a logged-in browser task once — replay it forever at zero LLM tokens. Local-first browser-automation MCP for Claude Code, Cursor & any MCP host; credentials never leave your machine.
- ✓Open-source license (MIT)
- ✓Actively maintained (<30d)
- ✓Clear description
- ✓Topics declared
claude mcp add tap -- npx -y @taprun/cli{
"mcpServers": {
"tap": {
"command": "npx",
"args": ["-y", "@taprun/cli"]
}
}
}MCP Servers overview
<p align="center">
<img src="extension/icons/icon.svg" width="120" height="120" alt="Taprun">
</p>
<h1 align="center">Taprun</h1>
<h4 align="center">
The action layer your agent drives in your own logged-in browser — for the exception-heavy, compliance-critical last 20% that clean APIs can't reach. Compiled once, replayed forever at zero LLM tokens.
</h4>
<p align="center">
<a href="https://taprun.dev/?utm_source=readme&utm_medium=docs&utm_campaign=homepage"><b>Homepage</b></a> |
<a href="https://taprun.dev/blog/?utm_source=readme&utm_medium=docs&utm_campaign=blog"><b>Blog</b></a> |
<a href="https://github.com/LeonTing1010/tap-skills"><b>Verified Claims</b></a> |
<a href="README.zh-CN.md"><b>中文</b></a>
</p>
<p align="center">
<a href="https://github.com/LeonTing1010/tap/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/LeonTing1010/tap/ci.yml?style=flat-square&label=CI" alt="CI"></a>
<a href="https://github.com/LeonTing1010/tap/releases/latest"><img src="https://img.shields.io/github/v/release/LeonTing1010/tap?style=flat-square" alt="Release"></a>
<a href="https://github.com/LeonTing1010/tap/stargazers"><img src="https://img.shields.io/github/stars/LeonTing1010/tap?style=flat-square" alt="Stars"></a>
<a href="LICENSE"><img src="https://img.shields.io/github/license/LeonTing1010/tap?style=flat-square" alt="License"></a>
<a href="https://chromewebstore.google.com/detail/tap/llcidejeoobdegbkolbjhfoeckphldce"><img src="https://img.shields.io/chrome-web-store/v/llcidejeoobdegbkolbjhfoeckphldce?style=flat-square&label=Chrome%20Web%20Store" alt="Chrome Web Store"></a>
</p>
<p align="center">
<img src="docs/assets/hero.svg" width="720" alt="Taprun: AI compiles a Hacker News tap once for $0.42, then replays forever at $0 per run">
</p>
---
**Your agent's closed-loop action layer for the browser — the login-gated, exception-heavy, compliance-critical last 20% that clean APIs and cloud agents can't reach. Driven in your own Chrome, captured once, replayed forever at zero LLM tokens.**
As APIs get walled off and metered, the work that survives lives behind logins, OTP walls, and human-gesture gates — the exceptions, approvals, and compliance steps a cloud agent architecturally can't touch. Taprun is the action layer for exactly that: your agent drives your real, already-logged-in Chrome, closes the loop (act → verify the effect → re-run on drift), and hands you a deterministic replay you own.
Every other browser agent re-runs a live LLM — and re-burns tokens — on every execution. Taprun's AI agent inspects the page **once** and emits a deterministic `.plan.json` program; every replay after that is pure data dispatch — same result every call, **$0 in tokens, no agent in the loop**. It runs in your real Chrome, so cookies and login sessions stay on your machine by architecture. `tap verify` catches breakage before your data goes stale.
Works with Claude Code, CodeBuddy, Cursor, Cline, Windsurf, and any MCP host — install straight from the chat window. Forge a tap from any URL on demand — no catalog needed.
```
Capture: AI inspects the site → compiles a .plan.json program (one-time cost)
Run: The program executes instantly, same result every time ($0, zero AI)
Verify: tap verify checks the snapshot equivalence predicate (catches drift)
Repair: re-run capture against the same site/name; the next (only when needed)
verify rebaselines after human review
```
## How Taprun Compares
| | Taprun | AI Browser Agents | Traditional Scrapers |
|--|-----|-------------------|---------------------|
| **AI cost per run** | $0 (compile once) | Tokens every run | Free |
| **Accuracy** | Deterministic | Varies per run | Deterministic |
| **Silent failure detection** | Per-tap CEL `snapshot_equivalent` predicate + 4-arm verdict | None | None |
| **Breakage diagnostics** | `tap verify` — exact diff of what changed | None | Manual spot checks |
| **Detection risk** | Low (real browser sessions) | High | High |
| **Runtimes** | 2 (Chrome extension + Playwright) | 1 | 1 |
| **Code inspectable** | .plan.json — bare JSON, 18-op closed vocabulary, git diff | Black box / ephemeral | Fragile scripts |
| **MCP native** | Yes (authoring layer only — execution is zero tokens) | No | No |
## Get Started
### 1. Attach to your agent — from the chat window
**Claude Code / CodeBuddy** — paste two lines into the chat, nothing else:
```
/plugin marketplace add LeonTing1010/taprun
/plugin install tap@taprun
```
That installs the Tap MCP server **plus** the skills that teach your agent when to use it and the hook that routes walled fetches to Tap — no terminal, no config file. (CodeBuddy wires plugin MCP servers at startup only, so **fully restart it once** after installing; Claude Code picks them up with `/reload-plugins`.)
**Any other MCP host** (Cursor · VS Code · Claude Desktop) — one command writes the config for you:
```bash
npx -y @taprun/cli embed cursor # or: vscode | claude-desktop | claude-code | codebuddy | qwen
```
The binary self-copies to `~/.tap/bin` and your agent's MCP config is written. Re-check anytime with `tap embed --verify`.
> **Using a coding agent that isn't in that list?** `tap embed` targets are *data, not code*: drop a row into `~/.tap/embed-targets.json` and the new agent works immediately — no engine release. Each row names one of four install *kinds* (`cc-plugin` for Claude-Code-plugin hosts, `cli-mcp-add` for CLIs with a `<cli> mcp add`, `ide-deeplink`, `desktop-bundle`), e.g. `[{"id":"kode","kind":"cc-plugin","display":"Kode CLI","tier":1,"cli":"kode"}]`. Same doctrine as the rest of Tap — the engine stays closed and mechanical; you extend it in local data.
Now pick your runtime — **the extension is only needed to reuse your *live* logged-in Chrome:**
- **Public pages / open APIs / CI — nothing more to install.** The MCP server runs over `npx`; you're done. Append `--no-extension` for a fully in-chat Playwright runtime with its own isolated profile (no browser gesture, no click).
- **Logged-in sites** (your bank / internal dashboard / Xiaohongshu / Zhihu) — just tell your agent **"set up tap for logged-in sites"** right in the chat. The **tap-setup** skill drives the whole bridge from the chat: it materializes the stable binary (from the engine `npx` already downloaded — no second download) and registers the native-messaging manifest, then opens the extension page. The single **[Add to Chrome](https://chromewebstore.google.com/detail/tap/llcidejeoobdegbkolbjhfoeckphldce)** click is the only step that isn't a chat action — it *is* the trust gate that lets Tap reuse your existing login, and the in-flight call resumes automatically once it lands.
- **Claude Desktop**: download [`tap.mcpb`](https://github.com/LeonTing1010/tap/releases/latest) and double-click.
<details>
<summary><b>Other install paths</b> (brew · curl · manual MCP JSON · raw binaries)</summary>
```bash
brew install LeonTing1010/tap/taprun # Homebrew (macOS / Linux)
curl -fsSL https://taprun.dev/install.sh | sh # permanent binary
npx -y @taprun/cli --version # zero-install (any Node host)
```
Manual MCP config, if you'd rather write it yourself:
```json
{ "mcpServers": { "tap": { "command": "npx", "args": ["-y", "@taprun/cli", "mcp", "stdio"] } } }
```
| Platform | Download |
|----------|----------|
| macOS (Apple Silicon) | [tap-macos-arm64](https://github.com/LeonTing1010/tap/releases/latest) |
| macOS (Intel) | [tap-macos-x64](https://github.com/LeonTing1010/tap/releases/latest) |
| Linux | [tap-linux-x64](https://github.com/LeonTing1010/tap/releases/latest) |
| Windows | [tap-windows-x64.exe](https://github.com/LeonTing1010/tap/releases/latest) |
</details>
### 2. Prove it works (~2 minutes, no login)
Run the first entry of the [claims ledger](https://github.com/LeonTing1010/tap-skills) — the exact verification its nightly CI runs:
```bash
mkdir -p ~/.tap/plans/github
curl -fsSL https://raw.githubusercontent.com/LeonTing1010/tap-skills/main/claims/2026-07-11-github-trending-has-no-api/plan.json \
-o ~/.tap/plans/github/trending-no-api.plan.json
tap github/trending-no-api
```
`"state": "committed"` plus today's trending repos at zero tokens = your install works **and** the claim holds.
### 3. Forge your own
```bash
tap capture https://news.ycombinator.com hn/front --intent "front-page stories with points"
tap hn/front # replay forever, $0
```
Or just ask your AI agent:
```
You: What's trending on GitHub today?
Agent: Here are today's top repos — React compiler hit 734 stars...
You: Capture a tap for Douban top 250 movies
Agent: Done. Run `tap douban/top250` anytime — $0 per run.
```
### Optional: Drive the binary from your own code (TypeScript / Python)
Skip MCP — call the `tap` binary from your own loop:
```bash
tap capture <url> hackernews/top --intent "front-page top stories"
tap hackernews/top --args '{}' # JSON-on-stdout, exit 0 on success
tap verify hackernews/top # 3-arm verdict (live / drifted / unreachable)
```
The CLI emits `ToolResult<T>` envelopes as JSON — same shape the MCP surface returns — so any language with a subprocess library can drive it. See `tap --help` for the full verb list.
### Have an existing Playwright / Puppeteer / Stagehand script?
Don't rewrite. Convert with one of the open-source adapters — drop your existing source in, get a Tap-compatible `.plan.json` plan out:
```bash
# Existing Playwright script (47M weekly npm downloads — most likely the one you have)
npm install @taprun/from-playwright @taprun/spec
node -e "import('@taprun/from-playwright').then(m => console.log(m.playwrightToTap(require('fs').readFileSync('tests/login.spec.ts','utf8'), {site:'example', name:'login'})))"
# Or scaffold a new starter from scratch
npx create-tap-script github/trending https://github.com/trending
```
| Adapter | SoWhat people ask about tap
What is LeonTing1010/tap?
+
LeonTing1010/tap is mcp servers for the Claude AI ecosystem. Capture a logged-in browser task once — replay it forever at zero LLM tokens. Local-first browser-automation MCP for Claude Code, Cursor & any MCP host; credentials never leave your machine. It has 12 GitHub stars and was last updated today.
How do I install tap?
+
You can install tap by cloning the repository (https://github.com/LeonTing1010/tap) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is LeonTing1010/tap safe to use?
+
Our security agent has analyzed LeonTing1010/tap and assigned a Trust Score of 87/100 (tier: Trusted). See the full breakdown of passed checks and flags on this page.
Who maintains LeonTing1010/tap?
+
LeonTing1010/tap is maintained by LeonTing1010. The last recorded GitHub activity is from today, with 2 open issues.
Are there alternatives to tap?
+
Yes. On ClaudeWave you can browse similar mcp servers at /categories/mcp, sorted by popularity or recent activity.
Deploy tap to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/leonting1010-tap)<a href="https://claudewave.com/repo/leonting1010-tap"><img src="https://claudewave.com/api/badge/leonting1010-tap" alt="Featured on ClaudeWave: LeonTing1010/tap" width="320" height="64" /></a>More MCP Servers
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
An open-source AI agent that brings the power of Gemini directly into your terminal.
The fastest path to AI-powered full stack observability, even for lean teams.
🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl!
Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface