---
id: speedict
version: "0.3.12"
license: unclear
license_treatment: permissive
maintenance: dormant
---

# speedict — Speedb Python Binding

License: permissive · Maintenance: dormant · Downloads: 145.8K/mo

## What it is and what it does

SpeeDict wraps RocksDB as a persistent, on-disk key-value store for Python. It operates in two modes: default mode (which serializes arbitrary Python objects like lists, dicts, NumPy arrays, and Pandas DataFrames using Pickle) and raw mode (which stores only bytes). The library provides a dictionary-like interface—you can set, get, iterate, and batch-retrieve keys—and persists data across sessions. It's built in Rust and compiled to wheels for common platforms.

The package is designed for applications that need fast, embedded key-value storage without running a separate database server. It supports column families, snapshots, batch writes, and RocksDB configuration options like write buffer size and compaction settings. However, it does not support merge operations or custom comparators, and maintenance has been dormant, meaning bug fixes and dependency updates are not expected.

Use it for:

- Cache frequently accessed Python objects (DataFrames, arrays) on disk without serializing to JSON or CSV.
- Build a lightweight embedded database for a single-machine application that needs persistent key-value storage.
- Store and retrieve raw bytes efficiently in raw mode for binary data or custom serialization schemes.
- Prototype or test database-backed logic without setting up a separate database server.
- Batch-load or batch-write large sets of key-value pairs with transactional semantics.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

SpeeDict is a Python key-value storage library backed by RocksDB that persists arbitrary Python objects (via Pickle) or raw bytes to disk, supporting dictionary-like access with batch operations.

Yes, if you need a fast, embedded key-value store for Python objects and can tolerate dormant maintenance. The library is stable for read-heavy workloads and has no known vulnerabilities. However, avoid it for mission-critical systems requiring active support or frequent updates. Verify that prebuilt wheels exist for your Python version and platform before committing.

## Install

pip install speedict
uv add speedict
poetry add speedict

## Installing speedict

Before you install: Medium install friction due to compiled wheels; prebuilt binaries available for Python 3.7, 3.10, 3.11, and 3.12 on macOS, Linux, and Windows, but no pure-Python fallback. Last release was 1007 days ago and maintenance status is dormant, so dependency updates and bug fixes are unlikely.

License in practice: Licensed under MIT (permissive), so you can use, modify, and distribute SpeeDict freely in commercial and private projects without restriction.

Quickstart:

from speedict import Rdict

db = Rdict("./my_store")
db["key"] = [1, 2, 3]
db[b"bytes_key"] = b"value"
print(db["key"])
db.close()

Requires a compiled extension; installation relies on prebuilt wheels for your platform and Python version.

Verify before relying:

- Whether dormant maintenance status affects stability or security for production use.
- Performance characteristics compared to other embedded key-value databases.
- Whether Pickle serialization overhead is acceptable for your workload.

## Package facts

- License: not declared (permissive)
- Python support: unspecified
- Install friction: medium
- Maintenance: dormant
- Downloads: 145.8K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

key-value store python, persistent dictionary, rocksdb python binding, on-disk key-value storage, python object serialization database, embedded database python, fast key-value storage, embedded-database, rocksdb-binding, persistence

[View on SkillFed](https://skillfed.io/packages/speedict) · [View on PyPI](https://pypi.org/project/speedict/)