Public app catalog

This catalog maps the public AGILAB app names users see in PROJECT to the package or source status behind them. It is intentionally more exhaustive than the README: the README stays focused on the shortest adoption path, while this page is the reference for choosing an app.

Before creating a new project, run:

agilab projects suggest "classic sklearn model with metrics"
agilab reuse suggest --from-notebook path/to/notebook.ipynb --kind project

The same reuse catalog also covers analysis views, so notebook-first work can discover both an existing app project and an existing visualization before it turns into a new maintained surface.

Package route legend:

PyPI app package: selected by the current release plan for standalone agi-app-* distribution. Installation requires a published package; entries awaiting their first publication are marked below.
Release artifact: selected by the release plan without standalone PyPI promotion. Published payload locations are listed below.
Source built-in: present in the public source checkout for development, examples, or compatibility, without a separate promoted payload package.

Project	Package	Package route	When to use it
`flight_telemetry_project`	`agi-app-flight-telemetry`	PyPI app package	First proof and compact end-to-end data ingestion with map/network analysis; also the real-world worker-only Cython example for a haversine speed kernel with reducer evidence.
`weather_forecast_project`	`agi-app-weather-forecast`	PyPI app package	Notebook-to-app migration using a small weather forecast dataset, forecast metrics, and promotion evidence. Historical `weather_forecast_legacy_project` identifiers resolve here when the app and AGILAB core come from the same release; new work should use the maintained name.
`sklearn_pipeline_project`	`agi-app-sklearn-pipeline`	PyPI app package	Minimal classic ML app proof using scikit-learn: deterministic dataset, fitted pipeline, serialized model, predictions, metrics, and artifact hashes.
`data_quality_gate_project`	`agi-app-data-quality-gate`	PyPI app package	Production-adjacent data gate for contract, drift, leakage, and promotion-decision evidence before training.
`mission_decision_project`	`agi-app-mission-decision`	PyPI app package	Deterministic mission-decision evidence with scenario scoring, re-planning, and decision artifacts.
`execution_pandas_project`	`agi-app-pandas-execution`	PyPI app package	Cython worker speedup demo for the Pandas path: typed contiguous `float64` kernel, Python/Cython runtime evidence, and reducer evidence.
`execution_polars_project`	`agi-app-polars-execution`	PyPI app package	Synthetic worker execution playground for the Polars path, comparable with the Pandas app.
`multi_app_dag_project`	`agi-app-multi-dag`	PyPI app package	Cross-app DAG template preview and artifact-handoff contract review.
`pytorch_playground_project`	`agi-app-pytorch-playground`	PyPI app package	Reproducible PyTorch classifier playground with live play/pause training, persisted controls, replayable teaching route, multi-UI surface declarations, evidence artifacts, and generated PyTorch/Lightning code. It can also be launched with `agilab pytorch-playground`, `agilab pytorch-playground --backend hf`, or the generic `agilab app surface pytorch_playground_project --ui <backend>` path.
`learning_assessment_project`	`agi-app-learning-assessment`	PyPI app package	Available on PyPI at version `2026.10.7`. Learning & Assessment provides evidence-scored diagnostic and self-evaluation cases, including the original TeSciA collection, with 2026 math coverage, a 12-case 2026 data-scientist interview evaluation spanning modern ML, RAG, agents, LLM evaluation, uncertainty, and token-cost optimization, classroom batch intake, live teacher dashboard artifacts, better-fix selection, and regression-plan evidence.
`uav_relay_queue_project`	`agi-app-uav-relay-queue`	PyPI app package	Compact UAV relay queue scenario with relay-health, scenario-cockpit, and network-map analysis.
`uav_queue_project`	`agi-app-uav-queue`	Release artifact	Queue-policy proof generator and scenario-cockpit evidence source used by the advanced proof pack.
`minimal_app_project`	`agi-app-minimal`	Release artifact	Minimal app structure reference for adapting manager, worker, settings, and app argument form code. Included in the published `agi-apps` package.
`r_runtime_bridge_project`	`agi-app-r-runtime-bridge`	Release artifact	Python manager and worker adapter for external R computations. Requires an installed external `Rscript` runtime.

All fourteen built-in projects are available from the published release. Eleven are promoted to PyPI through agilab[examples]. The Minimal App is included in agi-apps. The R Runtime Bridge and legacy UAV Queue source payloads are available in the archive of the signed v2026.10.07 tag. Their Python and scientific dependencies, and the bridge’s external Rscript runtime, must be installed separately.

Recommended first choices

Start with flight_telemetry_project when you want the shortest proof that AGILAB can install, execute, write evidence, and open analysis.
Use weather_forecast_project when the question is notebook migration and reusable forecast artifacts.
Use sklearn_pipeline_project when the question is classic ML reproducibility with a familiar scikit-learn pipeline.
Use pytorch_playground_project when you want a visual AI/ML demo that goes beyond a browser-only neural playground: live play/pause training, replay tokens, generated PyTorch/Lightning code, and exportable evidence.
Use uav_relay_queue_project or uav_queue_project only after the first proof, because they are better suited to advanced evidence and scenario comparison.

Package truth source

The PyPI publication status comes from the generated release plan and the package split contract, not from this page by hand. If the release plan changes, update this catalog in the same documentation pass.