dlthub-client
dlthub-client connects the open source data load tool (dlt) to the dltHub Runtime,
where you deploy, run and monitor pipelines, notebooks and apps. It contains:
dlthubCLI commands:login,deploy,run,serve,job,workspace,variable,dashboardand more, added to thedlthubcommand as a plugin.- MCP tools that let a coding agent inspect jobs, runs, logs, configuration and pipeline telemetry, served by
dlthub ai mcp. dlthub_sdk, a typed Python SDK, blocking andasyncio, for scripts, CI and services that call the platform.
Note: The Python SDK is in public preview. Its interface is stable and will only be extended, but its packaging may still change: it can move to another PyPI package and another top-level package name than
dlthub_sdk.
Installation
dlthub-client supports Python 3.11 and above.
For the CLI and the MCP tools, install dlt with the hub extra. It brings the dlthub command and a matching
dlthub-client:
pip install "dlt[hub]"
For the SDK alone, dlthub-client is enough. The SDK does not depend on dlt:
pip install dlthub-client
CLI
The Runtime commands appear inside a dlthub workspace. Create one, log in and deploy:
dlthub init
dlthub login
dlthub deploy
dlthub run <script> deploys and runs a batch job, dlthub serve <notebook> serves an interactive one, and
dlthub job and dlthub workspace manage what is deployed. Every command is listed in the
CLI reference.
MCP tools
The tools are opt-in features of the dlthub MCP server, which a workspace created with dlthub init comes with.
Run it with the features you need:
dlthub ai mcp run --features=+dlthub.jobs,+dlthub.logs
The features are dlthub.jobs, dlthub.logs, dlthub.config, dlthub.telemetry and dlthub.restore_pipeline.
dlthub ai mcp install --agent claude (or cursor, codex) writes the server into your agent's configuration.
Python SDK
Create an API key in the dltHub web app: a personal key (dlt_u_) in your account settings,
or a workspace key (dlt_sa_) in the workspace settings. Connect with it, then navigate from the runtime to the objects you need:
import dlthub_sdk
runtime = dlthub_sdk.connect(token="dlt_u_...")
ws = runtime.workspaces.get(id="...")
for job in ws.jobs.list():
print(job.job_ref, job.is_paused)
for result in ws.jobs.trigger(refs=["jobs.my_pipeline"]):
if result.started:
run = ws.job_runs.get(id=result.job_run_id).wait()
print(run.status)
connect_async returns the same surface for asyncio: calls to the platform are awaited and listings are async
iterables.
runtime = dlthub_sdk.connect_async(token="dlt_u_...")
ws = await runtime.workspaces.get(id="...")
async for job in ws.jobs.list():
print(job.job_ref)
Every failure raises a subclass of dlthub_sdk.DlthubError (NotAuthenticated, NotFound, NotAuthorized,
Conflict, ServerError, ...). When the platform sends a stable code, branch on it rather than on the message. For a token that rotates, pass credentials= instead of
token=. The SDK reads it before each request and renews it once if the platform rejects it.
The package is fully typed (py.typed). See the Python SDK reference
for every object and method.
Documentation
- CLI reference: every
dlthubcommand and its options. - Python SDK reference: the public
dlthub_sdkAPI. - REST API reference: the HTTP API the CLI and SDK are built on.
- dltHub docs: concepts and guides for the platform.
Metadata
Release files for dlthub-client 0.28.8
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| dlthub_client-0.28.8.tar.gz | 318.2 kB | Without provenance | Details for dlthub_client-0.28.8.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| dlthub_client-0.28.8-py3-none-any.whl | Python 3 | none | any | Details for dlthub_client-0.28.8-py3-none-any.whl |
Total release size: 855.5 kB
Release files / dlthub_client-0.28.8.tar.gz
| Download URL | dlthub_client-0.28.8.tar.gz |
|---|---|
| Size | 318.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e9e0d0634dbab4dc45ec846536ce62f5b177a91d9bd979d47a777ccd796856a4
|
|
BLAKE2b-256 checksum How to use checksums |
37aef38e68739526ca1a47ab726c8b1eb74d3ce573a6b0cf91d481013f2629ad
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.9.7
|
Release files / dlthub_client-0.28.8-py3-none-any.whl
| Download URL | dlthub_client-0.28.8-py3-none-any.whl |
|---|---|
| Size | 537.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
39c73038799cd9105b198dbe984b659f2be50613be9c3f58388097f3b811e1ac
|
|
BLAKE2b-256 checksum How to use checksums |
322c0084c394e638b81cc9c889b5ef8399bb0f7ddb9a306bff35e28c66bafb19
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.9.7
|
Security for dlthub-client 0.28.8
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.