A practical Python learning hub
Clear learning paths, first-party playbooks, and cited references for Python foundations, AI agents, automation, and production applications.
No account required · Primary sources first · Review dates shown
Action First
No fluff, no vague tutorials. Copy exact commands to reproduce tests locally, or provide structured rules directly to your AI IDE.
Clone the open-source repository and run the test suite to experience verifiable engineering with modern uv tooling.
git clone https://github.com/flypythoncom/python.git
cd python && uv run pytestPaste engineering guardrails directly into your .cursorrules, CLAUDE.md, or system prompt to prevent hallucinated changes.
curl -fsSL https://raw.githubusercontent.com/flypythoncom/python/master/templates/AGENT_RULES.example.mdLet your AI agent ingest curated Python playbooks, API boundaries, and catalogs without noisy HTML web scraping.
curl -fsSL https://flypython.com/llms.txtChoose a direction
Start where your current work is blocked. Each path links one guide to an actionable playbook, official references, and a named output.
Create an isolated project, make one function contract explicit, and verify success and failure behavior.
02Define one HTTP endpoint, its validation rules, dependency deadlines, test matrix, and release evidence.
03Write the run contract for a repeated task: checkpoints, idempotency, retry limits, and completion evidence.
04Run a deterministic agent loop, inspect its action-name trace, and test its boundaries before adding a model.
Featured
Run a no-key example, inspect each boundary, and test its failure paths before choosing a larger stack.
Download the Python 3.12 example, run it locally, verify its tested failure boundaries, and see where a real model belongs.
Read the guide →result = run_agent(
Task("My API request times out"),
max_steps=3,
)
for action in result.trace:
print(action)
print(result.answer)First-party playbooks
The four playbooks below open each learning path; the full set of twelve also covers API changes, migrations, releases, logging, and LLM evals. Official references remain available as supporting material.
A script works once, but its environment and failure behavior are not reproducible.
Python foundationsChecklist02A Python function needs an HTTP interface, but its validation, failures, and dependency limits are still implicit.
Web & APIsChecklist03A repeated task can restart or partially fail, but there is no durable definition of what completed.
AutomationChecklist04A task may need model judgment, but the action contract and failure boundaries have not been tested independently.
AI agentsChecklistHow FlyPython publishes
Published under the FlyPython Editorial organizational byline. Each guide exposes its sources, supported environment, update date, and either a runnable command or a named documentation-review environment and date.
Readers can rerun published code checks, inspect cited sources, and report reproducible corrections through the public issue tracker.