# Token launches (/scripting-v2/sources/launches)



Use `launches@ascend` or `launches@montra` to receive new Elysium mainnet launches while your script runs. Declare both selectors to watch both venues. The feed starts when the connection opens. Historical launches and events missed during a disconnect are outside the live feed.

```python
script = {"name": "launch-radar", "version": "2", "lookback": 20}
ASCEND = "launches@ascend"
MONTRA = "launches@montra"


def on_data(ctx, history):
    ascend = history.source(ASCEND, 0)
    montra = history.source(MONTRA, 0)
    launch = ascend if ctx.source == ASCEND else montra
    if launch is None:
        return
    return {"metrics": {"event": "launch", **launch}}
```

Save it as `launch-radar.py` and run `mlab script run launch-radar.py --duration 3600`. Follow its output with `mlab script logs JOB_ID --output jsonl --follow`, using the job ID printed by `script run`. Standalone CLI scripts use Furge's public indexed feed. Furge scripts use the feed on their assigned node.

The `metrics` from a real Montra launch look like this. The script adds `event` to the launch record so you can filter the logs.

```json
{
  "event": "launch",
  "exchange": "montra",
  "market": "99901:0xa3c935b0fdc1ff9453a2a0814e7d9fa7d05d1348",
  "quote": "0x0000000000000000000000000000000000000000",
  "quote_symbol": "HYPE",
  "created_at": 1794396371000
}
```

Use `market` to subscribe to the same token and get its symbol, `ser`, and available statistics. [See the matching market info example](/scripting-v2/sources/info#subscribe-during-a-run).

Both selectors use the same launch record schema.

| Field                  | Value                                                              |
| ---------------------- | ------------------------------------------------------------------ |
| `exchange`             | `ascend` or `montra`.                                              |
| `market`               | Canonical market key in the form `99901:<token-address>`.          |
| `quote`                | Quote asset address. Montra uses the zero address for native HYPE. |
| `quote_symbol`         | Display symbol of the quote asset.                                 |
| `created_at`           | Block time in Unix milliseconds.                                   |
| `buy_restricted_until` | Optional Montra early-buy fee window end in Unix milliseconds.     |

In Furge, choose **New script → Launch radar**, **Ascend sniper**, or **Montra sniper**. The sniper templates start in observe mode. Turn on `buy_enabled` and set a HYPE budget to queue one buy per run for a recent HYPE-pair launch. The Montra template skips launches with an active early-buy fee window.

These templates watch new launches across the selected venue. They do not arm the selected token in the Memes desk.

For a launch-driven buy, declare the venues your script can trade in `script["execution_venues"]`, then pass the launch's exchange and market to `ctx.buy`. When you omit `key`, the runtime deduplicates the action by its verified launch and side. To make an additional intended order on that market, supply a separate key.

```python
ctx.buy({
    "exchange": launch["exchange"],
    "market": launch["market"],
    "amount": "0.05",
    "slippage_bps": 100,
    "max_gas": "0.001",
})
```

Use [market info](/scripting-v2/sources/info) to read the token symbol, market cap, liquidity and holder data when available. If your script used `launch["symbol"]`, read the symbol from that feed. If it built an order key from transaction fields, let the runtime derive the key. Furge-managed script swaps on Ascend and Montra include a 2% Furge fee in addition to the pool fee. Use `on_execution` and `ctx.swap_fill()` to read confirmed settlement.
