remarkapy

A Python package for interacting with the reMarkable Cloud API.

Quickstart

from remarkapy import Client

client = Client()
items = client.list_items()
for item in items[:5]:
    print(item.type, item.visibleName, item.parent, item.id, item.hash)

Use client.list_items() for cheap library scans and lookup workflows. client.list_hydrated_items() is the explicit full-hydration variant and fetches each item's manifest, metadata, and .content blob.

Download a raw bundle:

from remarkapy import Client

client = Client()
first = client.list_items()[0]
client.download_raw_bundle(first.hash, "bundle.zip")

Upload a PDF:

from remarkapy import Client

client = Client()
with open("example.pdf", "rb") as handle:
    created = client.put_pdf("Example.pdf", handle.read())
print(created)

CLI examples:

uv run remarkapy init
uv run rkpy ls
uv run rkpy ls Papers/
uv run rkpy ls -r Papers/
uv run rkpy get Papers/Example.pdf
uv run rkpy get Papers/Example.pdf --output ./Example.pdf
uv run rkpy get "Meeting Notes" --format bundle
uv run rkpy export "Meeting Notes" ./exports --format pdf
uv run rkpy info Papers/Example.pdf
uv run remarkapy get-id <item-id-or-hash>
uv run rkpy put-pdf ./example.pdf --parent Papers/
uv run rkpy mkdir Inbox
uv run rkpy rename Papers/Example.pdf "New Name"
uv run rkpy move Papers/Example.pdf Inbox/
uv run rkpy trash Papers/Old Draft.pdf

Why another

As far as I can tell, the major reMarkable client SDKs have all suffered from API churn.

• rmapy is archived.
• rmapi is archived (but resurrected??)
• reMarkable continues to change endpoints and backend behavior.

remarkapy aims to stay useful by:

• discovering the current public webapp/raw hosts at runtime,
• supporting the current immutable-manifest sync protocol,
• exposing a clean Python library API first,
• keeping real integration tests available behind an explicit opt-in flag.

Optional export backends

• remarkapy does not render annotations itself.
• Install the remarks command separately if you want export support (for example: uv tool install git+https://github.com/avncharlie/remarks.git).
• rkpy export shells out to that installed remarks executable by default, but --remarks-cmd still lets you override it.
• For single-file PDF/Markdown exports, you can pass an exact output file path like ./notes.pdf; otherwise pass an output directory.
• remarks needs the system Cairo library for PDF rendering; on macOS install it with brew install cairo.
• Export output is staged in a temporary bundle and copied back into your requested output directory.

Features

• Base authentication and device-token reuse
• Endpoint discovery
• List documents and folders
• Download original PDFs and EPUBs
• Download raw item bundles as zip archives
• Upload PDFs and EPUBs
• Create folders
• Rename items
• Move items
• Delete items by moving them to trash
• Mock-backed pytest suite
• Read-only live-account smoke tests

Roadmap

• Download annotated PDFs rendered by the service, if available
• Add a small CLI wrapper for common operations
• Add safer live mutation tests using temporary fixtures
• Add sync-to-directory and sync-from-directory helpers
• Add device registration from a dedicated command-line flow

Testing

Dangerzone Safety

There are live tests in this repo that interact with the real reMarkable Cloud API. These tests are marked with the live pytest marker and are skipped by default to prevent accidental mutations of real accounts.

The live tests in this repo are intentionally read-only by default.

To run them, opt in explicitly:

REMARKAPY_RUN_LIVE=1 uv run pytest -q -m live

The destructive paths (rename, move, delete, put_*) are covered by the mock test suite. Before adding live destructive tests, use temporary folders or dummy uploads that are safe to clean up.

Injected or mock-backed clients should use in-memory tokens or an explicit test config path; they must not fall back to the default rmapi config location.