Developer Ecosystem

Headless simulation that fits your existing pipeline

Trigger runs from your own servers, batch-process massive datasets, and pull rendered output straight into your tools — REST, GraphQL, and native SDKs.

Browse the node library

Official SDKs

Native wrappers for Python, Julia & R

Interact with Poly Sim OS directly from Jupyter without ever opening the web interface. The same graph object powers the canvas and the SDK — one schema, every surface.

Generate graphs from prompts in code
Run local-hint or cloud-escalated jobs
Stream results into pandas / DataFrames
Export CSV, JSON, and MP4 artifacts

import polysimos as ps

sim = ps.Client(api_key="ps_live_...")

# Prompt → validated graph
graph = sim.ai.generate(
    "double pendulum under chaotic gravity"
)
graph.nodes["solver"].params["dt"] = 0.001

job = sim.simulate(graph, compute="auto")   # local hint / cloud escalate
df = job.result().to_dataframe()            # straight into pandas
print(df.head())

Core endpoints

A REST surface for everything, with a GraphQL gateway for shaped queries across jobs, nodes, and renders.

POST/v1/simulateRun a headless simulation from a graph; returns a job_id.

GET/v1/jobs/{id}Poll status, progress, and token cost for a run.

POST/v1/ai/generate-graphTurn a prompt into a validated node graph.

GET/v1/nodes/libraryFetch standard and community logic blocks.

POST/v1/nodes/publishPublish a node cluster to the community library.

GET/v1/render/exportDownload rendered output as CSV, JSON, or MP4.

Compute Token Model

Deterministic, metered pricing

A run's token cost is computed server-side from degrees of freedom $N$ , time steps $T$ , and a solver weight $w_{s}$ . No surprises — you can estimate it before you submit.

tokens = ⌈ \frac{w _{s} \cdot N \cdot T}{1 0 ^{9}} ⌉

cloud cost model · billed on job completion

w_s: explicit 1.0 · implicit/stiff 2.4 · FEM 3.1

API tier limits & throughput
Plan	Rate limit	API calls/mo	Concurrent jobs
Independent Researcher	20 req/s	10,000	2
Lab / Agency	100 req/s	unlimited	16
Institution	400 req/s	unlimited	64
Enterprise	custom	unlimited	dedicated

REST + GraphQL

Trigger headless runs, batch datasets, and pull output into any pipeline.

Community Library API

Fetch, publish, and fork node clusters programmatically as your registry grows.

Compute Tokens

Cloud jobs are metered server-side. Estimate cost client-side; debit on completion.

See API limits by plan