Download Changefiles

Download incremental updates to keep your local data current

Changefiles let you download only what’s changed since your last snapshot, so you can keep a local copy of OpenAlex up to date without re-downloading the full dataset. Each day’s changefile contains every entity that was created or modified on that date. The API keeps the last 60 days of changefiles available.

List Available Dates

The /changefiles endpoint returns all available changefile dates:

curl "https://api.openalex.org/changefiles?api_key=YOUR_KEY"

{
  "meta": {
    "count": 9
  },
  "results": [
    {
      "date": "2026-02-20",
      "url": "https://api.openalex.org/changefiles/2026-02-20?api_key=YOUR_KEY"
    },
    {
      "date": "2026-02-19",
      "url": "https://api.openalex.org/changefiles/2026-02-19?api_key=YOUR_KEY"
    },
    {
      "date": "2026-02-18",
      "url": "https://api.openalex.org/changefiles/2026-02-18?api_key=YOUR_KEY"
    }
  ]
}

Get a Day’s Changes

Follow a date’s URL to see which entities changed or were created - and how many of each - and download links in both JSONL and Parquet:

curl "https://api.openalex.org/changefiles/2026-02-19?api_key=YOUR_KEY"

{
  "meta": {
    "count": 19,
    "date": "2026-02-19"
  },
  "results": [
    {
      "entity": "works",
      "records": 3171968,
      "formats": {
        "jsonl": {
          "size_bytes": 10842055680,
          "size_display": "10.1 GB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/works_2026-02-19.jsonl.gz?api_key=YOUR_KEY"
        },
        "parquet": {
          "size_bytes": 12130148352,
          "size_display": "11.3 GB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/works_2026-02-19.parquet?api_key=YOUR_KEY"
        }
      }
    },
    {
      "entity": "authors",
      "records": 3124690,
      "formats": {
        "jsonl": {
          "size_bytes": 9664790758,
          "size_display": "9.0 GB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/authors_2026-02-19.jsonl.gz?api_key=YOUR_KEY"
        },
        "parquet": {
          "size_bytes": 6938735872,
          "size_display": "6.5 GB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/authors_2026-02-19.parquet?api_key=YOUR_KEY"
        }
      }
    },
    {
      "entity": "institutions",
      "records": 48857,
      "formats": {
        "jsonl": {
          "size_bytes": 95621065,
          "size_display": "91.2 MB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/institutions_2026-02-19.jsonl.gz?api_key=YOUR_KEY"
        },
        "parquet": {
          "size_bytes": 40703645,
          "size_display": "38.8 MB",
          "url": "https://content.openalex.org/changefiles/2026-02-19/institutions_2026-02-19.parquet?api_key=YOUR_KEY"
        }
      }
    }
  ]
}

Each entry includes:

• entity — the entity type (works, authors, institutions, etc.)
• records — total number of created or modified records
• formats — download URLs and file sizes for JSONL (gzipped) and Parquet

What’s in a Changefile

A day’s changefile contains every entity record that was created or modified on that date. This includes:

• Newly created entities (e.g., a paper published today)
• Existing entities with updated metadata (e.g., a work that gained new citations)

Each record is a complete entity object — the same format you’d get from the API. To apply the update, upsert into your local data using the entity’s id as the primary key.

Incremental Update Workflow

1. Note the date of your last full snapshot or changefile download
2. List available changefiles and identify dates after your last update
3. For each new date, download the entity files you need
4. Upsert records into your local copy using id as the primary key

import requests

API_KEY = "YOUR_KEY"

# 1. List available changefiles
dates = requests.get(
    f"https://api.openalex.org/changefiles?api_key={API_KEY}"
).json()["results"]

# 2. Pick dates after your last update
last_update = "2026-02-17"
new_dates = [d for d in dates if d["date"] > last_update]

# 3. Download each day's changes
for date_info in new_dates:
    day = requests.get(date_info["url"]).json()

    for entity in day["results"]:
        if entity["entity"] == "works":
            download_url = entity["formats"]["jsonl"]["url"]
            print(f"{date_info['date']}: {entity['records']:,} works "
                  f"({entity['formats']['jsonl']['size_display']})")
            # download and upsert...

Formats

Changefiles are available in two formats:

Format	Extension	Best for
JSONL (gzipped)	.jsonl.gz	Streaming ingestion, line-by-line processing
Parquet	.parquet	Analytics, loading into data warehouses (BigQuery, Spark, DuckDB)

Both contain the same data — choose whichever fits your pipeline.