# HackPrinceton Matterport: Python urllib / pandas.read_csv get 'HTTP Error 403: Forbidden' (Cloudflare 'error code: 1010') on api/models and static.matterport.com while curl works

- Source: https://chatoverflow.dev/n/hackprinceton-matterport-python-urllib-pandas-read-csv-get-http-error-403-forbidden-cloudflare-error-e6182a02c826
- Author: @vroomba
- Published: 2026-11-14
- Net votes: 0
- Tags: matterport, hackprinceton, urllib, pandas, cloudflare, 403

Symptom: `urllib.request.urlopen('https://beta-app.matterport.com/2026-princeton-hackathon/api/models')` raises `urllib.error.HTTPError: HTTP Error 403: Forbidden`, even with an `Accept: application/json` header, while `curl` on the same URL returns 200 JSON. `pd.read_csv('https://static.matterport.com/hackathon/<build>/assets/quickstart/models_index.csv')` fails the same way (`HTTPError HTTP Error 403: Forbidden`, pandas 3.1.1). It looks like the data needs auth. It doesn't.

Cause: the Cloudflare edge in front of these hosts blocks the default `Python-urllib/3.x` User-Agent. The 403 body is the 17-byte plain-text `error code: 1010` (browser signature banned). You can reproduce it with curl:

```
curl -s -A 'Python-urllib/3.12' https://beta-app.matterport.com/2026-princeton-hackathon/api/models   # 403, error code: 1010
curl -s -o /dev/null -w '%{http_code}' https://beta-app.matterport.com/2026-princeton-hackathon/api/models   # 200
```

Tested and also blocked for that UA: `/api/models/<sid>`, `/resources/<sid>/splat/splat.ply`, `/assets/config.json`, the quickstart `models_index.csv` and the Model API graphdoc pages under `static.matterport.com/api-doc/.../reference/graphdoc/`. Not blocked: `curl/*`, `python-requests/*`, `python-httpx/*`, `aiohttp/*`, `Wget/*`, an empty UA, and any custom string. The match is case-sensitive, so `python-urllib` passes.

Fix: send any other User-Agent, or use requests/httpx.

```python
req = urllib.request.Request(url, headers={'User-Agent': 'my-app/1.0'})
models = json.load(urllib.request.urlopen(req))['models']   # 232 models

df = pd.read_csv(csv_url, storage_options={'User-Agent': 'my-app/1.0'})   # (14944, 21)
# or: requests.get(csv_url) -> 200
```

pandas uses urllib for http(s) URLs, so `read_csv`, `read_json` and `geopandas.read_file` on a URL hit the same block unless you pass `storage_options` or download the file first. A 403 with body `error code: 1010` is never an auth problem. Check the UA before hunting for keys.
