Skip to main content

DuDuClaw Python SDK

Python companion package for DuDuClaw — the Multi-Agent AI Assistant Platform.

This package carries the Python-side helpers that sit next to the DuDuClaw Rust binary:

  • duduclaw.mcp — MCP server helpers: API-key auth with scope enforcement, memory tools (store / read / search / namespace / quota)
  • duduclaw.agents — capability-based agent routing (manifest loader, matcher, router, memory resolver)
  • duduclaw.evolution — evolution vetter (GVU self-play verification)
  • duduclaw.tools — agent tool definitions

Installation

pip install duduclaw

Note: This package is a companion to the main DuDuClaw binary. Install the binary via npm (npx duduclaw) or the install script — see the repository README.

Requirements

  • Python 3.10+
  • The anthropic and httpx packages (installed automatically)

Memory evaluation harness (repository only)

python/duduclaw/memory_eval/ is a benchmarking harness for memory retrieval quality. It is not part of the published wheel — no runtime path calls it, and shipping 7k lines of benchmark code to every pip install bought nobody anything. Run it from a checkout instead:

git clone https://github.com/zhixuli0406/DuDuClaw && cd DuDuClaw
# The harness needs four packages the published wheel does not depend on:
pip install aiohttp asyncpg datasets pytest pytest-asyncio aioresponses
PYTHONPATH=python python -m duduclaw.memory_eval.smoke_test         # ~5 min P0 smoke run
PYTHONPATH=python pytest python/duduclaw/memory_eval/tests -v       # the harness's own tests

Without asyncpg, aiohttp, and datasets the modules will not even import — one more reason they have no business in a wheel whose declared dependencies are anthropic, httpx, and pyyaml.

About its data: data/golden_qa_set.jsonl holds 200 hand-written question/answer pairs (every row is tagged source: "manual"). It is a DuDuClaw-authored zh-TW set, not LOCOMO and not derived from it, despite what some older internal notes said. fetch_benchmarks.py can pull LongMemEval and PersonaMem separately; those downloads are not bundled either.

License

Apache-2.0

Metadata

Release files for duduclaw 1.75.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for duduclaw 1.75.0
File Size Uploaded Provenance Actions
duduclaw-1.75.0.tar.gz 51.8 kB Without provenance Details for duduclaw-1.75.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for duduclaw 1.75.0
File Interpreter ABI Platform Actions
duduclaw-1.75.0-py3-none-any.whl Python 3 none any Details for duduclaw-1.75.0-py3-none-any.whl

Total release size: 113.3 kB

Release files / duduclaw-1.75.0.tar.gz

Download URL duduclaw-1.75.0.tar.gz
Size 51.8 kB
Tags Source
SHA-256 checksum
How to use checksums
67b4a7db238477563ce2ea73f1f0cf0bc5cb1357d2b6568470c827192e0e0360
BLAKE2b-256 checksum
How to use checksums
01e94fa513891424338d4567dbb2f808cf4cb5a740d998dfb5629f4e4b4ec664
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.14

Release files / duduclaw-1.75.0-py3-none-any.whl

Download URL duduclaw-1.75.0-py3-none-any.whl
Size 61.5 kB
Tags Python 3
SHA-256 checksum
How to use checksums
d88b9d5ea3cc892dde077223173d7123b6785f346febd74e0dfffd94ff184f8c
BLAKE2b-256 checksum
How to use checksums
4f5d50b108b2bcfb9acdd70eb46d4fb11f3f825684d5b8620f63b0729ced7d11
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.14

Security for duduclaw 1.75.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

1.75.0 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page