AxFlow For Python

This skill helps an agent write Python code with the generated Ax package axllm. Use the generated package API, examples, and manifests; do not import TypeScript-only APIs unless you are editing the TypeScript package.

When To Use

• Compose generators, agents, and nested flows into a workflow graph.
• Reason about flow state, node inputs, returns, caching, and errors.
• Use generated package examples for flow graphs and provider-backed flows.

Package Facts

• Language: Python.
• Package: axllm.
• Package API docs: API.md and axir-api.json.
• Capability manifest: axir-capabilities.json.
• Runnable examples: examples/.
• Real network support: yes.
• Scripted no-key transport support: yes.
• Runtime profiles: javascript-quickjs, python-pyodide.

Core Pattern

from axllm import ax, flow

draft = ax("topicText:string -> draftText:string")
wf = (
    flow({"id": "docs.coreFlow"})
    .execute("draft", draft, {"reads": ["topicText"], "writes": ["draftResult", "draftText"]})
    .returns({"draftText": "draftText"})
)

More Patterns

Typed programs

Build each flow node from its own input/output contract.

classifier = ax('requestText:string -> route:class "support, sales, engineering"')
responder = ax("requestText:string, route:string -> responseText:string")

Class decision

Declare reads and writes so the responder waits for the typed route.

branch_flow = (
    flow({"id": "docs.branchFlow"})
    .execute("classifier", classifier, {"reads": ["requestText"], "writes": ["classifierResult", "route"]})
    .execute("responder", responder, {"reads": ["requestText", "route"], "writes": ["responderResult", "responseText"]})
    .returns({"route": "route", "responseText": "responseText"})
)

Parallel fan-out and join

Independent reads let research and audience analysis share one planner group.

parallel_flow = (
    flow({"id": "docs.parallelFlow"})
    .execute("research", research, {"reads": ["topicText"], "writes": ["researchResult", "factList"]})
    .execute("audience", audience, {"reads": ["topicText"], "writes": ["audienceResult", "audienceAngle"]})
    .execute("join", join, {"reads": ["factList", "audienceAngle"], "writes": ["joinResult", "briefText"]})
    .returns({"briefText": "briefText"})
)

Draft, critique, revise

A linear refinement pipeline makes each dependency explicit.

refine_flow = (
    flow({"id": "docs.refineFlow"})
    .execute("draft", draft, {"reads": ["topicText"], "writes": ["draftResult", "draftText"]})
    .execute("critique", critique, {"reads": ["draftText"], "writes": ["critiqueResult", "critiqueText"]})
    .execute("revise", revise, {"reads": ["draftText", "critiqueText"], "writes": ["reviseResult", "revisedText"]})
    .returns({"revisedText": "revisedText"})
)

Run a flow

Forward accepts the provider client and the public flow inputs.

output = parallel_flow.forward(client, {"topicText": "Typed LLM workflows"})

Start from the complete programs under examples/, then browse the larger gallery at https://axllm.dev/python/subsystems/flow/.

Relevant API Surface

• Flow: flow, AxFlow

Guardrails

• Start from package examples for exact native syntax before inventing a new call shape.
• Use provider-api examples only when the user explicitly has provider credentials available.
• Use no-key examples for deterministic local checks and provider request mapping.
• Treat AxIR as the source of generated package truth: if package docs disagree with source code, update the compiler and regenerate packages.
• Do not copy repo-maintainer skills from tools/*/skills/ into user packages.