dex-screener-scanner

Files (4)

• SKILL.md
• dexscreener_layout.md · references/
• scan_dexscreener.py · scripts/
• screen_tokens.py · scripts/

SKILL.md (first lines)

# DexScreener Solana Scanner

Autonomous agent skill for scanning and screening Solana tokens on DexScreener using browser automation (Playwright/Puppeteer). The goal is to navigate `dexscreener.com/solana`, extract token data from live listings, apply screening filters, and identify the best trading opportunities.

## Overview

This skill uses browser automation to:

1. Navigate to DexScreener's Solana board
2. Scrape live token listings (new, trending, gainers, etc.)
3. Extract key metrics for each token
4. Apply configurable screening filters to find the best opportunities
5. Return ranked results with rationale

## Workflow

### Step 1: Launch Browser & Navigate

Use a browser automation tool (Playwright or Puppeteer) to open `https://dexscreener.com/solana`.

```python
# Example with Playwright
page = await browser.new_page()
await page.goto("https://dexscreener.com/solana", wait_until="domcontentloaded")
```

**Important:** DexScreener is a heavy React SPA. Wait for the token table rows to render before scraping. The table typically appears within 3-8 seconds.

### Step 2: Wait for Token Table to Load

The token listing is a dynamic table. Watch for:

- CSS selector for table rows: `div[data-group="token-row"]` or `.ds-table-row` or `tr[data-id]`
- The table has multiple tabs: **New**, **Trending**, **Gainers**, **Recently Copied**, etc.
- Wait for at least one row to be visible before scraping

Recommended wait strategy:

```python
# Wait for first token row to appear
await page.wait_for_selector('[data-group="token-row"]', timeout=15000)
# Or wait for the table body
await page.wait_for_selector('table tbody tr', timeout=15000)
```

If the wait times out, the page may be loading slowly — retry once with a longer timeout.

### Step 3: Parse Token Listing Data

Each token row typically contains these fields (column order may vary slightly):

| Field | Description | Example |
|-------|-------------|---------|
| **#** | Rank/position | 1, 2, 3 |
| **Name/Symbol** | Token name + ticker | "dogwifhat / WIF" |
| **Price** | Current SOL or USD price | "$0.0042" |
| **Price Change** | 5m/1h/6h/24h change % | "+15.3%" |
| **Volume** | 24h trading volume | "$1.2M" |
| **Liquidity** | Pool liquidity | "$45K" |
| **Market Cap** | Fully diluted market cap | "$2.1M" |
| **Age** | How long since creation | "3m" (3 minutes), "1h" |
| **Txns** | Transaction count (buys/sells) | "1.2K / 800" |
| **Holders** | Unique holder count | "450" |
| **FDV / Liq** | FDV-to-Liquidity ratio | "12.5x" |

**Parsing approach:**

```python
rows = await page.query_selector_all('[data-group="token-row"]')
tokens = []
for row in rows:
    cells = await row.query_selector_all('td')
    # Extract based on cell index
    tokens.append({
        "rank": await cells[0].inner_text(),
        "name": await cells[1].inner_text(),
        "price": await cells[2].inner_text(),
        "change_5m": await cells[3].inner_text(),
        "volume": await cells[4].inner_text(),
        "liquidity": await cells[5].inner_text(),
        "market_cap": await cells[6].inner_text(),
        "age": await cells[7].inner_text(),
        "txns": await cells[8].inner_text(),
        "holders": await cells[9].inner_text(),
        # ... etc
    })
```

**Note:** Cell indices may shift if DexScreener updates their layout. Always inspect the actual DOM first. Use `page.evaluate()` for more robust scraping if query selectors are unreliable.

### Step 4: Switch Between Listing Tabs

DexScreener has multiple tabs that reveal different token sets. Click these to scan more broadly:

```python
# Available tabs (selectors may vary):
# New: Most recently created pairs
await page.click('button:has-text("New")')
# Trending: Hottest tokens right now
await page.click('button:has-text("Trending")')
# Gainers: Top gainers
await page.click('button:has-text("Gainers")')
# Recently Copied: Tokens being copied from other chains
await page.click('button:has-text("Recently Copied")')
```

Wait 2-3 seconds after clicking a tab for the table to re-render.

### Step 5: Scroll to Load More Tokens

DexScreener loads tokens in batches (about 25 per page). Scroll down to trigger lazy loading:

```python
# Scroll to bottom of table to load more
for _ in range(3):
    await page.evaluate('window.scrollBy(0, 800)')
    await page.wait_for_timeout(1500)
```

From x402agent/solana-clawd, MIT.

More from x402agent/solana-clawd