https://github.com/ancs21/playpanda
Fetch any URL to clean, LLM-ready markdown. 3-tier engine: HTTP → Lightpanda → CloakBrowser.
https://github.com/ancs21/playpanda
ai cli cloakbrowser headless-browser lightpanda llm markdown scraper web-scraping zig
Last synced: 28 days ago
JSON representation
Fetch any URL to clean, LLM-ready markdown. 3-tier engine: HTTP → Lightpanda → CloakBrowser.
- Host: GitHub
- URL: https://github.com/ancs21/playpanda
- Owner: ancs21
- License: apache-2.0
- Created: 2026-03-28T16:24:47.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2026-04-25T09:42:43.000Z (about 2 months ago)
- Last Synced: 2026-05-24T01:34:14.201Z (28 days ago)
- Topics: ai, cli, cloakbrowser, headless-browser, lightpanda, llm, markdown, scraper, web-scraping, zig
- Language: Zig
- Size: 3.5 MB
- Stars: 9
- Watchers: 0
- Forks: 1
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
- Security: SECURITY.md
Awesome Lists containing this project
README
```
╭───────────╮
│ ■ ■ │
│ ▶ │
│ ╰──╯ │
╰───────────╯
```
# playpanda
[](https://github.com/ancs21/playpanda/actions/workflows/ci.yml)
[](https://github.com/ancs21/playpanda/releases/latest)
[](LICENSE)
[](https://ziglang.org)
[](https://github.com/ancs21/playpanda/releases/latest)
Fetch any webpage as clean, LLM-ready markdown. Single binary, zero config.
## Why playpanda?
| | playpanda | Crawl4AI | Firecrawl | Jina Reader |
|---|:---:|:---:|:---:|:---:|
| Single binary | Zig | Python + pip | Node + API key | curl only |
| Zero config | Yes | Nearly | No | Yes |
| Auth built-in | Login + cookies | Hooks/profiles | Dashboard | x-set-cookie |
| Anti-bot | 3-tier auto-escalation | Manual config | Partial | No |
| API key required | No | No | Yes | Freemium |
| Avg speed | 1.8s | ~3s | ~5s | ~2s |
| Avg tokens/page | 2,868 | varies | varies | varies |
## How It Works
playpanda uses a 3-tier fetch engine that automatically escalates until it gets content:
| Tier | Method | Speed | When |
|------|--------|-------|------|
| 1 | HTTP + native Zig HTML-to-markdown | ~150ms | Default for most sites |
| 2 | [Lightpanda](https://lightpanda.io/) headless browser | ~1.5s | When Tier 1 returns empty/broken content |
| 3 | [CloakBrowser](https://github.com/CloakHQ/CloakBrowser) | ~6s | Bot-protected sites (Facebook, LinkedIn, Medium, etc.) |
If a page is blocked or empty, playpanda automatically tries the next tier.
## Install
One-liner (installs binary + all dependencies):
```
curl -fsSL https://raw.githubusercontent.com/ancs21/playpanda/main/scripts/install.sh | sh
```
### From source
Requires [Zig](https://ziglang.org/) 0.15+:
```
git clone https://github.com/ancs21/playpanda.git
cd playpanda
zig build -Doptimize=.ReleaseFast
cp zig-out/bin/playpanda ~/.local/bin/
```
### Use as Agent Skill
```
npx skills add ancs21/playpanda
```
### Dependencies
The installer handles these automatically, or install manually:
- [Lightpanda](https://lightpanda.io/) — headless browser for Tier 2
- Python 3 + `websockets` — Tier 3 stealth browser and cookie harvesting
- Chrome/Chromium — for Tier 3 stealth and login flow
## Usage
### Fetch a page
```
playpanda https://example.com # markdown to stdout
playpanda https://example.com > article.md # save to file
```
### Fetch multiple pages
```
playpanda https://example.com,https://ziglang.org # multiple URLs, separated by ---
```
### Log in to sites
```
playpanda profile # opens browser, log in, press Enter
```
Cookies are saved to `~/.playpanda/cookies.json` and used automatically on subsequent fetches.
### Upgrade
```
playpanda upgrade
```
## How Cookies Work
1. **`playpanda profile`** opens a browser with CDP enabled. Log in to any site, then press Enter. Cookies are harvested via Chrome DevTools Protocol and saved.
2. **Fetch flow**: Cookies are automatically loaded and matched by domain. Tier 1 uses HTTP cookie headers, Tier 2 injects via `Network.setCookies` over CDP.
## Bot-Protected Sites
These domains automatically use Tier 3 (stealth browser):
facebook.com, instagram.com, linkedin.com, x.com, twitter.com, medium.com, google.com, youtube.com, tiktok.com, reddit.com, substack.com, threads.net, pinterest.com
Other sites start at Tier 1 and escalate if blocked.
## License
[Apache 2.0](LICENSE)