ClarosHQ Python SDK — provider-neutral AI observability and tracing client (OpenAI, Anthropic, LangChain adapters)
Project description
claroshq — ClarosHQ Python SDK
claroshq is the official Python SDK for the ClarosHQ AI observability and
governance platform. The core client is provider-neutral and
model-neutral; provider-specific instrumentations (OpenAI, Anthropic Claude,
LangChain) are shipped as optional extras and normalize events into a
generic ClarosHQ trace and span schema.
Installation
Install the pinned 0.1.0 release from production PyPI:
python -m pip install claroshq==0.1.0
ClarosHQ Sprint 2 (S2-019) gates the public release on a clean-machine install proof against the same hosted index that external customers use. Until that proof has been recorded against production PyPI, the command above is the intended install path — the production upload of
0.1.0is performed by thePublish SDKworkflow on asdk/v0.1.0tag push and is gated by Omar's release decision. See docs/sprint-2/s2-019-builder-package.md for the current status and the release-target decision.
Install from TestPyPI (staging only)
TestPyPI is the staging index — it is used for the final dry-run of
upload, metadata rendering, and fresh-environment install before the
package is promoted to production PyPI. TestPyPI does not mirror
production PyPI, so the --extra-index-url https://pypi.org/simple/ flag
is required to resolve claroshq's runtime dependency httpx>=0.27 and
any transitive dependencies:
python -m pip install claroshq==0.1.0 \
--index-url https://test.pypi.org/simple/ \
--extra-index-url https://pypi.org/simple/
A successful install from TestPyPI is a useful staging milestone, but it does not satisfy the ClarosHQ Sprint 1 commercial-completion gate — production PyPI is the release target for that.
Optional provider adapters
Optional provider adapters can be installed via extras (use the same
==0.1.0 pin to keep your install reproducible until you intentionally
move to a newer release):
python -m pip install "claroshq[anthropic]==0.1.0"
python -m pip install "claroshq[langchain]==0.1.0"
python -m pip install "claroshq[all]==0.1.0" # anthropic + langchain + openai
Note: there is no dedicated
openaiextra —openaiis bundled withclaroshq[all]. If you only need the OpenAI adapter, install the OpenAI client directly alongsideclaroshq(e.g.pip install claroshq==0.1.0 openai).
Editable install (contributors)
If you are working inside the ClarosHQ platform monorepo, install the SDK in editable mode with the development extras:
python -m venv sdk/.venv
sdk/.venv/Scripts/python -m pip install -U pip # Windows
# source sdk/.venv/bin/activate # macOS / Linux
sdk/.venv/Scripts/python -m pip install -e "./sdk[dev]"
The dev extra installs pytest, ruff, black, mypy, pre-commit,
build, and twine — everything needed to develop, lint, test, build, and
publish the package.
Verify the install
After installing, confirm the SDK loaded correctly and reports the expected version:
python -c "import claroshq; print(claroshq.__version__)"
This should print:
0.1.0
Quickstart
import claroshq
claroshq.init(
api_key="YOUR_CLAROSHQ_INGEST_KEY",
endpoint="https://ingest.claroshq.ai",
)
# Auto-instrument every supported provider that is importable in this process:
claroshq.instrument_all()
# ... your existing OpenAI / Anthropic / LangChain code runs unchanged ...
claroshq.flush()
See the main repository README for the full local development and backend setup.
License
Apache-2.0. See the project metadata in pyproject.toml.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file claroshq-0.1.0.tar.gz.
File metadata
- Download URL: claroshq-0.1.0.tar.gz
- Upload date:
- Size: 15.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
99cb41eae7795ee87956c69f4ee7d08607e883f0392450fc3075608ff6864f82
|
|
| MD5 |
1745a6ae2ab667d7b519d69aecebd0b5
|
|
| BLAKE2b-256 |
86cfa722b9f728b6021dfc0f9911ebcbd8051c51d8064811b85f085e0f625adb
|
File details
Details for the file claroshq-0.1.0-py3-none-any.whl.
File metadata
- Download URL: claroshq-0.1.0-py3-none-any.whl
- Upload date:
- Size: 20.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
6b577e04452f9a1250c3ae704e83fd81542931a467a01206ba63b555a87dce93
|
|
| MD5 |
52bce23ef54ece3e2b8c95333b8eb77b
|
|
| BLAKE2b-256 |
fba9449b7122765c61ee0152e79d9ddcb7b5cd9c40a5efc4b6a8e12700bf38fe
|