nanoodle 0.1.1

Run nanoodle visual AI workflows from Python — zero-dependency executor for saved noodle-graph.json files

nanoodle (Python)

Run visual AI workflows from Python. Design them in the nanoodle editor, save as noodle-graph.json, then load and re-run them here — same graph, same NanoGPT API, your own key.

Zero runtime dependencies (stdlib only). Library + CLI in one install.

Looking for JavaScript / Node? → nanoodle-js

At a glance

Build once, run anywhere. The browser app is for designing and testing. This package is for automating the same workflows in scripts, servers, and agents.

Package	nanoodle on PyPI
Runtime	Python ≥ 3.9 · stdlib only · no deps
Sibling	JavaScript package (same graphs, same semantics)
Editor	nanoodle.io — wire nodes, hit 💾, download the graph

Install

pip install nanoodle
export NANOGPT_API_KEY=...   # nano-gpt.com API key (or OAuth access token)

Quickstart (library)

from nanoodle import Workflow

wf = Workflow.load("noodle-graph.json")
result = wf.run({"Text": "a cozy ramen shop on a rainy night"})
result["Image"].save("ramen.png")            # media: MediaRef (url + bytes()/save())
print(result.cost_usd, result.remaining_balance)

With the app’s starter graph (text → LLM prompt-writer → image), that’s the whole program.

Discover a workflow’s interface

wf.inputs    # [InputSpec(key="Text", node_id="n1", field="text", kind="textarea", ...)]
wf.outputs   # [OutputSpec(key="Image", node_id="n3", type="image", ports=["image"])]
wf.settings  # [SettingSpec(key="n3.size", kind="select", default="1k", ...)]

Input keys are flexible (case-insensitive): the node’s custom name, nodeId.field ("n2.system"), or the input’s label when unique. A workflow with exactly one required input also accepts a bare value: wf.run("hello").

Media inputs

from nanoodle import media_from_file

wf.run({"Image": media_from_file("photo.jpg")})            # local file
wf.run({"Image": "https://example.com/photo.jpg"})         # hosted or data: URL
wf.run({"Image": raw_bytes})                               # raw bytes (MIME sniffed)

Media is sent inline as base64 (NanoGPT has no upload endpoint). Files over ~4.4 MB (~3.5 MB for transcription) are refused locally with a clear error before any paid call.

Settings, progress, errors

result = wf.run(
    {"Text": "sunset harbor"},
    settings={"n3.model": "flux-dev", "n3.size": "1k"},
    timeout=600,
    on_progress=lambda evt: print(evt["type"], evt.get("name", "")),
)

run() raises RunError when an output (sink) node fails — error.result still has partial results, per-node statuses, and cost so far. Failures in lanes no output depends on only appear in result.errors. Unknown/unsupported node types, missing required inputs, bad keys, and a missing API key all fail before anything is spent.

CLI

Installed as nanoodle-py (and python -m nanoodle always works):

nanoodle-py inspect graph.json
nanoodle-py run graph.json --input Text="a cozy ramen shop" --set n3.size=1k --out ./out
nanoodle-py run graph.json --input n2.system=@style.txt --json
nanoodle-py run graph.json --env-file .env --input Text="hello"   # NANOGPT_API_KEY from a .env file

• --out DIR — save media outputs to files
• --json — machine-readable result
• --env-file PATH — load .env-style KEY=VALUE lines (existing env vars win)

Supported nodes

runs	node types
local	text, upload (image/audio/video), choice, join, comment
NanoGPT	llm (incl. vision + audio input), image, draw, edit, inpaint*, vision, tvideo, ivideo, vedit, lipsync, music, remix, tts, transcribe
not supported (browser-only media processing)	resize, vframes, combine, soundtrack, trim, extractaudio

Workflows with unsupported node types load with a warning and fail fast at run() with UnsupportedNodeError — before any network call.

* inpaint: the browser app composites the mask onto black at the source pixel size; this library passes your mask through verbatim. Supply a black/white mask matching the source dimensions.

Use it as an agent skill

A saved workflow plus a short SKILL.md playbook is a skill any coding agent can run — Claude Code (.claude/skills/<name>/SKILL.md) or anything that reads markdown and runs shell. Recipe and template: docs/agent-skills.md. Full example: examples/agent-skill/poster-generator/.

Cost

Bring your own NanoGPT API key; NanoGPT bills your balance per generation and reports the price on each response.

• result.cost_usd — total of prices returned
• result.cost_exact — False if any call omitted a price (total is then a floor)
• result.remaining_balance — freshest balance the API reported

A price of 0 means known-included (subscription), not unknown. No telemetry, no analytics; the API key is never logged.

Testing

Tests run fully offline against a mock NanoGPT server (tests/harness/):

python -m unittest discover -s tests -t .

Opt-in live probe (spends a fraction of a cent): python3 scripts/live-spot-check.py (add --image to also run the starter graph’s image step).

Docs

Design contract and format/engine/io specs live in docs/: DESIGN.md, SPEC-format.md, SPEC-engine.md, SPEC-io.md.

Same contract as the JavaScript package.

License

MIT — see LICENSE. Not affiliated with NanoGPT. Build workflows at nanoodle.io.