Newcomer Guide
If you are new to AGILab, optimize for one outcome: one successful local run of a built-in app. Do not start with cluster installs, private app repositories, or packaging work.
This page gives the shortest mental model for the framework. For the exact installation commands, see Quick-Start.
What success looks like
You are “past the newcomer hurdle” when all of the following are true:
A local install/run completes without needing SSH or remote workers.
AGILab writes logs and generated snippets under
~/log/execute/<app>/.You can do that either from the web UI or from a notebook using
AgiEnvandAGI.run(...).
Choose your path
Pick one path before you start, but do not treat them as equal:
Recommended first proof path: clone the repository, use the web UI, and run the built-in
flight_projectlocally.Evaluate AGILab quickly: use the published package in a fresh
uv-managed virtual environment.Contribute or inspect the framework deeply: use the source checkout and then branch into notebook, wrappers, or cluster flows after the first local proof works.
If you are unsure, choose the recommended flight_project path from
Quick-Start. Mixing the packaged install and a source checkout too early
is one of the easiest ways to confuse paths, environments, and generated
scripts.
Prefer code first?
If you already live in Jupyter, use Notebook Quick Start for the fastest local notebook-first route. This is a supported newcomer path, not just an advanced API reference.
The first 10 minutes
Install or launch AGILab with the commands from Quick-Start.
Keep the first run local. Skip SSH hosts and cluster settings.
Start with
src/agilab/apps/builtin/flight_projectas the default first proof path. Usemycode_projectonly if you explicitly want the smallest reference app instead of the clearest product demonstration.Choose one first-run workflow:
Notebook-first: open Notebook Quick Start and run the local
AgiEnv+AGI.run(...)example from Jupyter.UI-first: use the AGILAB UI in this order: PROJECT -> ORCHESTRATE -> PIPELINE -> ANALYSIS.
Confirm that
~/log/execute/<app>/contains fresh output.
The core ideas, in plain language
App / project: a runnable AGILab workload with settings, manifests, and worker code.
ORCHESTRATE: the page that turns settings into install, distribute, and run actions.
PIPELINE: the place to save and replay generated steps instead of retyping code snippets.
ANALYSIS: Streamlit pages for post-run inspection.
Worker: the isolated runtime that actually executes the workload.
Common newcomer traps
Mixing package mode and source mode without being explicit: pick one first, then switch deliberately.
Trying cluster mode before a local run succeeds: local success gives you a clean baseline for later SSH debugging.
Expecting private or optional apps to appear automatically: public built-in apps live under
src/agilab/apps/builtin; extra apps usually requireAPPS_REPOSITORY/AGILAB_APPS_REPOSITORY.Running ``uvx agilab`` from the source tree: from a repository checkout, use the source commands documented in Quick-Start so you do not accidentally run the published wheel.
Contributors editing generated docs output: edit
docs/sourcein the docs repository, notdocs/html.
Where to go next
Quick-Start for the shortest install/run commands.
Notebook Quick Start for the code-first notebook route.
AGILab Architecture for the runtime control path.
ORCHESTRATE for the ORCHESTRATE workflow.
Distributed Workers after the local path works and you are ready for SSH or multi-node execution.