Dropbox

Read and write files in Dropbox directly from Fused UDFs using OAuth-managed tokens. No API keys are hardcoded — Fused refreshes the token on your behalf. Work with the official dropbox SDK client, or read files directly with dropbox:// paths via fsspec (pandas, pyarrow, etc.).

App Folder scope

The Fused Dropbox app uses Dropbox's App Folder permission, so every path is sandboxed to a single dedicated folder in your Dropbox (it appears as Apps/<AppName>/). All paths below — both SDK paths like /report.csv and dropbox://report.csv URLs — are relative to that app folder root; the rest of your Dropbox is not accessible.

Setup

1. Connect Dropbox

Open Integrations & Secrets in Workbench, find the Dropbox section, and click Connect Dropbox. You will be redirected to Dropbox to authorize access, then back to Workbench. You can also connect from the CLI with fused integrations dropbox connect.

2. OAuth scopes

When you connect, the following scopes are requested:

Scope	Description
files.metadata.read	List folders and read file/folder metadata
files.content.read	Download file contents
files.content.write	Upload, move, and delete files
account_info.read	Read the account email shown in the integrations UI

---

dropbox:// path format

dropbox://file.csv                    # a file in the app folder root
dropbox://folder/file.parquet         # a file inside a subfolder

All dropbox:// paths are relative to your Dropbox App Folder root — do not prefix them with Apps/<AppName>. The same paths work in the Workbench file explorer once connected.

---

Examples

Quick start

fused.api.dropbox_connect() returns a FusedDropboxConnection; call .client() to get an official dropbox.Dropbox client. The client is built with your refresh token and the app key, so the SDK refreshes access tokens itself — no app secret is exposed, and it keeps working for long-running jobs.

@fused.udf
def udf():
    import pandas as pd
    from dropbox.files import WriteMode

    dbx = fused.api.dropbox_connect().client()

    # Upload a file
    dbx.files_upload(b"hello,world\n", "/example.csv", mode=WriteMode("overwrite"))

    # Download a file
    metadata, response = dbx.files_download("/example.csv")
    print(response.content)

    # List the app folder (root is the empty string)
    result = dbx.files_list_folder("")
    return pd.DataFrame([{"name": e.name} for e in result.entries])

The .client() method requires the dropbox package (pip install dropbox), which is already available in the Fused execution environment. See the Dropbox Python SDK reference for the full API (files_*, sharing_*, users_*, etc.).

Read a file with dropbox:// paths

Fused registers a dropbox:// fsspec filesystem, so any library that accepts an fsspec URL can read from your Dropbox app folder directly:

@fused.udf
def udf():
    import pandas as pd
    return pd.read_csv("dropbox://data/measurements.csv")

@fused.udf
def udf():
    import pandas as pd
    return pd.read_parquet("dropbox://exports/cities.parquet")

This filesystem is read-only — use the SDK client to write.

Common operations

@fused.udf
def udf():
    import pandas as pd
    from dropbox.files import WriteMode

    dbx = fused.api.dropbox_connect().client()

    # List a subfolder (auto-paginate)
    entries = []
    res = dbx.files_list_folder("/reports")
    entries.extend(res.entries)
    while res.has_more:
        res = dbx.files_list_folder_continue(res.cursor)
        entries.extend(res.entries)

    # Create a folder
    dbx.files_create_folder_v2("/reports/2026")

    # Upload (overwrite if it exists)
    dbx.files_upload(b"...bytes...", "/reports/2026/q1.csv", mode=WriteMode("overwrite"))

    # Download
    metadata, response = dbx.files_download("/reports/2026/q1.csv")
    data = response.content

    # Get a temporary, directly-downloadable link (valid ~4h)
    link = dbx.files_get_temporary_link("/reports/2026/q1.csv").link

    # Delete
    dbx.files_delete_v2("/reports/2026/q1.csv")

    return pd.DataFrame([{"name": e.name} for e in entries])

Read input and write output

@fused.udf
def udf():
    import pandas as pd
    from dropbox.files import WriteMode

    # Read an input straight from Dropbox via fsspec
    df = pd.read_csv("dropbox://input/raw.csv")

    out = df.describe().reset_index()

    # Write a result back with the SDK client
    dbx = fused.api.dropbox_connect().client()
    dbx.files_upload(
        out.to_csv(index=False).encode(),
        "/output/summary.csv",
        mode=WriteMode("overwrite"),
    )
    return out

CLI

fused integrations dropbox connect   # print the OAuth URL (opens a browser in a tty)
fused integrations dropbox revoke    # disconnect and remove stored tokens
fused integrations list              # show connection status for all integrations

Disconnecting

To revoke the integration, go to Integrations & Secrets, find the Dropbox section, and click Disconnect (or run fused integrations dropbox revoke). This removes stored tokens from both the Fused server and Secrets Manager. Files already written to your Dropbox app folder are not affected.