← backapi docs

berryline api

read-only access to the same estimate, chart, and report data the website itself uses. free for personal/educational use — sign in with your harvard email to get a key. all endpoints below are versioned (v1), so they won't change shape under you.

authentication

every request below needs an Authorization header with your API key as a DRF token. get one on the account page — it's free, just sign in with a harvard.edu email.

Authorization: Token <your-api-key>

rate limit: 1000 requests/day per key. requests without a valid token get a 401.

GET/api/v1/status

current estimate

the live, recency-weighted estimate of how many people are in line right now.

request

curl -H "Authorization: Token <key>" \
  https://api.howlongistheberryline.com/api/v1/status

response

{
  "has_estimate": true,
  "message": null,
  "estimate_display": "15-20",
  "estimate_value": 17.3,
  "report_count": 4,
  "last_updated": "2026-08-29T14:32:10-04:00",
  "hours": {
    "is_open": true,
    "closing_soon": false,
    "minutes_until_close": 145,
    "closes_at": "10:50pm",
    "opens_at": null
  }
}

GET/api/v1/chart

chart data

scatter of individual reports plus a smoothed line series over a date range. omit both params for today.

query params

since — YYYY-MM-DD, inclusive, local time. optional.
until — YYYY-MM-DD, inclusive, local time. optional.

request

curl -H "Authorization: Token <key>" \
  "https://api.howlongistheberryline.com/api/v1/chart?since=2026-08-01&until=2026-08-28"

response

{
  "scatter": [
    { "t": "2026-08-28T14:05:00-04:00", "value": 15 }
  ],
  "line": [
    { "t": "2026-08-28T00:00:00-04:00", "value": null },
    { "t": "2026-08-28T00:15:00-04:00", "value": 12.4 }
  ]
}

GET/api/v1/patterns

typical patterns

plain average headcount by hour-of-day and by day-of-week, over an optional range. defaults to all-time. useful for 'when's it usually short' questions rather than a live number.

query params

since — YYYY-MM-DD, inclusive, local time. optional.
until — YYYY-MM-DD, inclusive, local time. optional.

request

curl -H "Authorization: Token <key>" \
  https://api.howlongistheberryline.com/api/v1/patterns

response

{
  "by_hour": [
    { "hour": 0, "value": null, "report_count": 0 },
    { "hour": 14, "value": 18.2, "report_count": 41 }
  ],
  "by_weekday": [
    { "weekday": 0, "label": "Mon", "value": 14.1, "report_count": 88 }
  ]
}

GET/api/v1/reports

recent reports feed

raw, individual reports, newest first, cursor-paginated. no lat/lng — those never leave the server.

query params

limit — page size, default 8, max 200.
cursor — opaque pagination cursor from a previous response's next field.

request

curl -H "Authorization: Token <key>" \
  "https://api.howlongistheberryline.com/api/v1/reports?limit=50"

response

{
  "next": "https://api.howlongistheberryline.com/api/v1/reports?cursor=abc123",
  "previous": null,
  "results": [
    { "id": 512, "bucket": "11-20", "created_at": "2026-08-29T14:32:10-04:00", "photo_url": null }
  ]
}

GET/api/v1/export.csv

csv export

every report in the range as a csv download — timestamp, bucket, whether a photo was attached. same privacy rule as the feed: no lat/lng.

query params

since — YYYY-MM-DD, inclusive, local time. optional.
until — YYYY-MM-DD, inclusive, local time. optional.

request

curl -H "Authorization: Token <key>" \
  "https://api.howlongistheberryline.com/api/v1/export.csv?since=2026-08-01" -o reports.csv

response

timestamp,bucket,has_photo
2026-08-01T09:14:02-04:00,6-10,False
2026-08-01T09:41:55-04:00,11-20,False

buckets

reports are collected as one of six headcount buckets: 0, 1-5, 6-10, 11-20, 20-30, 30+. chart/patterns/status endpoints blend these into a smoothed numeric estimate — that's a separate, continuous number, not one of the six buckets.

questions or need a higher rate limit? igzcodes@gmail.com