DEVELOPER & ENTERPRISE HUB

Drop to code. Scale to the cluster.

The canvas is never a cage. Break into Python or R at any node, wire your DevOps, and offload the heavy lifting to high-performance compute — without leaving the pipeline.

from k0gn1t import Canvas, op

pipe = Canvas("churn.k0")

data = pipe.add(op.ReadCSV, path="customers.csv")

@pipe.execute_python(inputs=[data])

def clean(df):

    df = df.dropna(subset=["age"])

    df["income"] = df["income"].str.replace(",", "").astype(float)

    return df

model = pipe.add(op.GradientBoostedTrees, n_trees=180)

pipe.connect(clean, model).run(offload="gpu-cluster")

Built for the way engineers ship

Version, test, and deploy pipelines with the tooling you already trust.

Git-native versioning

Every pipeline is a diffable file. Branch, review, and merge changes like any codebase.

CI/CD pipelines

Trigger runs from GitHub Actions, GitLab, or Jenkins. Gate deploys on model metrics.

Containerized runtime

Reproducible Docker images per pipeline — same result on laptop, CI, and prod.

REST & Python SDK

Call any pipeline as an endpoint or embed it in your app with a two-line SDK.

Observability

Structured logs, run traces, and drift alerts stream to Datadog or OpenTelemetry.

Export anywhere

One click compiles a canvas to standalone Python, R, or a portable ONNX model.

HIGH-PERFORMANCE COMPUTING

Offload the heavy nodes to dedicated compute

Flag any operator to run on GPU clusters, Spark, or your on-prem grid. k0gn1t schedules, parallelizes, and streams results back to the canvas — the graph never blocks.

✓GPU, Spark, and on-prem grid targets per node

✓Automatic parallelization & back-pressure

✓Results stream back without blocking the graph

cluster · churn.k0offloading

gpu-1

63%

gpu-2

84%

gpu-3

89%

gpu-4

74%

gpu-5

50%

gpu-6

36%

gpu-7

41%

gpu-8

63%

gpu-9

84%

gpu-10

89%

gpu-11

74%

gpu-12

50%

Enterprise-ready from day one

Security, governance, and control your platform team will sign off on.

⛨

SSO & SAML

Okta, Azure AD, and SCIM provisioning.

⌂

On-prem / VPC

Deploy in your own cloud or air-gapped.

≡

Audit & RBAC

Fine-grained roles and full audit trails.

◷

99.98% SLA

Dedicated support and uptime guarantees.

Ship your first pipeline to production

Start on the free tier, then bring SSO, on-prem, and dedicated compute when you're ready.