Copy-trade signals - 0xinsider API

Build a follow list of S- and A-grade wallets, compare their copy scores, and print recent large trades from those wallets. Each matching trade includes the token ID when 0xinsider has it. Every trade has already executed on Polymarket. Its price is the wallet’s fill price, so copying it later may produce a different result. The API provides analytics; your own Polymarket client handles any orders. You need a Pro API key and Python with requests installed. Run python -m pip install requests if needed. The snippets share the variables defined in step 1.

1. Build a follow list

Leaderboard lists S-, A-, and B-grade wallets, sorted by score. It has no min_grade filter, so keep S and A in your code.

import os
from datetime import datetime, timedelta, timezone

import requests

BASE = "https://api.0xinsider.com/api/v1"
H = {"Authorization": f"Bearer {os.environ['OXINSIDER_API_KEY']}"}

def get(path, **params):
    response = requests.get(f"{BASE}{path}", params=params, headers=H, timeout=30)
    response.raise_for_status()
    return response.json()


leaders = get("/leaderboard", limit=100)
follow = [t["address"] for t in leaders["data"] if t["grade"] in ("S", "A")]

This example reads the first 100 wallets. To include more, follow next_cursor with the same filters. If you receive cursor_expired, discard that traversal and start again from the first page.

Field or filter	Meaning
`grade`	The wallet’s grade from settled P&L relative to other tracked wallets. See Grades.
`realized_pnl`	P&L on closed positions, including fees and rebates.
`pnl`	P&L including open positions, which changes with market prices.
`strategy`	An optional filter for observable activity, such as `high_activity`. See Wallet trading styles.

2. Compare copy scores and freshness

Batch traders reads up to 25 wallets per call. Use expand: ["quant_metrics"] to request copy_score and smart_score.

Field	Meaning
`copy_score`	A score from 0 to 100 for how suitable the wallet’s history is to copy. It includes penalties for limited history, concentration, sizing, large losses, and inconsistent edge.
`smart_score`	The skill score before copy-specific penalties.
`quant_metrics`	An optional block. It is omitted when the stored metrics are at least 6 hours old.
`data_quality`	Field-group timestamps and availability. The `ranking` group dates the grade and scores.
Item `status`	`ok` or `error` for each input. One failed input does not fail the batch.

A fresh ranking group has a timestamp, but you still need to compare it with your own maximum age. The example accepts rankings no older than 6 hours. It skips undated rankings and sorts wallets without a copy_score last. Step 3 watches only wallets accepted by this freshness check.

ranked = []
for start in range(0, len(follow), 25):
    response = requests.post(
        f"{BASE}/traders/batch",
        json={"traders": follow[start:start + 25], "expand": ["quant_metrics"]},
        headers=H,
        timeout=30,
    )
    response.raise_for_status()
    for item in response.json()["data"]:
        if item["status"] != "ok":
            print("skip:", item["input"], item["error"]["message"])
            continue
        trader = item["data"]
        ranking = next(
            (group for group in trader["data_quality"]["field_groups"]
             if group["group"] == "ranking"), None
        )
        if ranking is None or ranking["status"] != "fresh" or "as_of" not in ranking:
            print("skip, grade has no usable timestamp:", trader["address"])
            continue
        as_of = datetime.fromisoformat(ranking["as_of"].replace("Z", "+00:00"))
        if datetime.now(timezone.utc) - as_of > timedelta(hours=6):
            print("skip, grade older than 6 hours:", trader["address"], ranking["as_of"])
            continue
        metrics = trader.get("quant_metrics")
        score = metrics.get("copy_score") if metrics is not None else None
        ranked.append((score, trader["address"], trader.get("username"), trader.get("grade")))

ranked.sort(key=lambda row: (row[0] is not None, row[0] if row[0] is not None else 0), reverse=True)
for score, address, username, grade in ranked:
    print(score, grade, username or address[:10])

A missing or null score means insufficient or unavailable data; it is not a low score. The quant metrics page defines the calculation. SDK users can apply their own freshness limit with assessDataQuality, assess_data_quality, or AssessDataQuality.

3. Match recent trades against the list

Large trades returns recent trades, newest first. It has no wallet filter, so match trader.address against your list.

follow_set = {address.lower() for _, address, _, _ in ranked}
seen = set()


def new_moves(min_size=10000):
    page = get("/large-trades", min_grade="A", min_size=min_size, limit=100)
    moves = []
    for tr in page["data"]:
        if tr["id"] in seen:
            continue
        seen.add(tr["id"])
        if tr["trader"]["address"].lower() in follow_set:
            moves.append(tr)
    return moves


for tr in new_moves():
    who = tr["trader"].get("username") or tr["trader"]["address"][:10]
    print(
        who, tr["side"], tr["outcome"], f'${tr["size_usd"]:,.0f}',
        "@", tr["price"], tr["market"]["title"], tr["token_id"],
    )

Field or filter	Meaning
`min_grade=A`	Keeps S- and A-grade wallets.
`min_size`	Sets the minimum USD trade value. The endpoint default is $5,000; this example uses$ 10,000.
`id`	Identifies the trade. Use it for deduplication.
`side`	`BUY` or `SELL`.
`outcome`	The outcome label. It is `null` for markets with more than 2 outcomes.
`token_id`	The outcome’s Polymarket CLOB token ID, or `null` when unavailable. Older recorded trades can also lack it.

This example reads one page and stores deduplication IDs in memory. It can miss trades during a long pause and repeats them after a restart. For a durable consumer, use the event replay recipe in Alert bot. To read one wallet’s historical large trades, use Large trades history with trader.

4. Inspect entries and exits in one market

Take market.condition_id from a matching trade and request Position timeline.

curl -H "Authorization: Bearer $OXINSIDER_API_KEY" \
  "https://api.0xinsider.com/api/v1/trader/0x204f72f35326db932158cba6adff0b9a1da95e14/position-timeline?condition_id=0x123..."

running_amount gives the share balance after each stored fill. running_avg_price gives the buy-weighted average price; selling does not change it. Both include earlier pages, so you do not need to recompute them while paging. The timeline excludes other position-changing operations, such as redemptions. It is available only for closely tracked wallets. A 404 with trader_not_tracked means the timeline is unavailable; retrying does not create it.

Limits

A wallet’s historical grade or score does not guarantee a future result.
The feed contains recorded large trades, not every fill from a followed wallet.
Trade prices describe executed fills, not quotes available to you.
The API does not place, cancel, or read orders.
The polling example does not provide durable delivery. Use Alert bot when missing an event is unacceptable.

Use Review a trader’s history before deciding which wallets to follow.

​1. Build a follow list

​2. Compare copy scores and freshness

​3. Match recent trades against the list

​4. Inspect entries and exits in one market

​Limits

1. Build a follow list

2. Compare copy scores and freshness

3. Match recent trades against the list

4. Inspect entries and exits in one market

Limits