Legacy V1 pagination

Use this page if your application already calls a V1 inventory endpoint. These routes remain supported for compatibility but are deprecated for new integrations. New applications should use Live Page, or Refresh → Status → Current for complete saved inventories.

Choose your existing route

Both return the original unwrapped inventory body on success. Neither returns V2's data and meta envelope. Keep the route your client already handles; switching paths can change its error parsing.

Make a request

Set STEAMCALL_API_KEY to your Service Account Credential secret in your local shell. Replace the example Steam ID with the public inventory you want to read.

curl --fail-with-body --max-time 15 \
  'https://api.steamcall.com/api/v1/inventory/76561198000000000/440/2?language=english' \
  --header "X-API-Key: $STEAMCALL_API_KEY"

• steam_id: the owner's 17-digit SteamID64.
• app_id: the Steam game ID; 440 is Team Fortress 2.
• context_id: the game's inventory context; TF2 player items normally use 2.
• language: description language, defaulting to english.
• start_assetid: omit on the first request. On subsequent requests, pass the previous response's last_assetid unchanged.

The page size is fixed at up to 2,000 assets. There is no page_size or opaque cursor parameter on these routes.

Understand continuation

A successful response uses top-level assets, descriptions, total_inventory_count, and success: 1. When another page exists, more_items is set and last_assetid supplies the continuation value. Keep asset IDs as strings; they can exceed JavaScript's safe integer range.

1. Collect the returned assets and descriptions.
2. If more_items is absent or zero, the traversal is complete.
3. Otherwise, send last_assetid as start_assetid with the same owner, game, context, and language.
4. Stop and mark the traversal incomplete if the continuation is missing, repeats, or the page/deadline limit is reached.

Inventories can change while you read them. A V1 traversal is not a pinned revision; use Current when you need pages from one saved revision.

Bounded Python example

This server-side example requires Python and the requests package. It requests at most 125 pages, checks a 120-second overall budget between requests, and gives each request at most 15 seconds of connect/read timeout. The HTTP timeout is not a strict wall-clock deadline for a streaming response. It does not automatically retry or publish partial results.

import os
import time
import requests


def read_legacy_inventory(steam_id, app_id=440, context_id=2):
    url = f"https://api.steamcall.com/api/v1/inventory/{steam_id}/{app_id}/{context_id}"
    headers = {"X-API-Key": os.environ["STEAMCALL_API_KEY"]}
    params = {"language": "english"}
    assets = {}
    descriptions = {}
    seen = set()
    deadline = time.monotonic() + 120

    for _ in range(125):
        remaining = deadline - time.monotonic()
        if remaining <= 0:
            raise RuntimeError("Inventory read exceeded its deadline; results are incomplete")
        response = requests.get(url, headers=headers, params=params, timeout=min(15, remaining))
        response.raise_for_status()
        body = response.json()
        if body.get("success") != 1:
            raise RuntimeError("Inventory page was not successful")
        for asset in body.get("assets", []):
            assets[asset["assetid"]] = asset
        for description in body.get("descriptions", []):
            key = (description["classid"], description["instanceid"])
            descriptions[key] = description
        if not body.get("more_items"):
            return {"assets": list(assets.values()), "descriptions": list(descriptions.values())}
        continuation = body.get("last_assetid")
        if not isinstance(continuation, str) or not continuation.isdecimal() or continuation in seen:
            raise RuntimeError("Missing or repeated continuation; results are incomplete")
        seen.add(continuation)
        params["start_assetid"] = continuation

    raise RuntimeError("Inventory page limit reached; results are incomplete")

Only replace your application's saved inventory after this function returns successfully. Preserve the previous data on failure; a private inventory is not an empty inventory.

Billing and errors

Each requested page uses the Live Page price, not one price for the full inventory. A confirmed private-inventory response is also billable under the active price policy. Read X-Credits-Charged and X-Credits-Remaining; the original route also retains X-Api-Credits-Remaining for older clients. See Usage and credits.

Keep the response's X-Request-ID for support. For exact schemas, open the original V1 reference or V1 alias reference.