---
id: snc4onnx
version: "1.0.14"
license: MIT License
license_treatment: permissive
maintenance: active
---

# snc4onnx — Simple tool to combine onnx models. Simple Network Combine Tool for ONNX.

License: permissive · Maintenance: active · Downloads: 84.3K/mo

## What it is and what it does

snc4onnx is a command-line and Python API tool for merging two or more ONNX neural network models into a single combined model. It works by connecting the output operators of one model to the input operators of another, with optional prefixing of operator names to prevent naming conflicts when models share common operation names. The tool runs ONNX simplification on the merged result by default to remove redundant operations.

The package is designed for developers who want to compose pre-trained models without writing custom graph manipulation code. It supports both file-based workflows (reading and writing .onnx files) and in-memory graph objects, making it suitable for both CLI automation and programmatic integration into larger pipelines.

Use it for:

- Combine encoder and decoder ONNX models into a single end-to-end inference model for deployment.
- Merge feature extraction and classification models trained separately into one unified network.
- Chain multiple specialized models (e.g., preprocessing, main inference, post-processing) into a single deployable artifact.
- Fuse stereo vision or multi-stage models where outputs of one stage feed into the next.
- Automate model composition in CI/CD pipelines without writing custom ONNX graph code.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Merges multiple ONNX neural network models into a single combined model by connecting specified output and input operators, with optional operator name prefixing to avoid conflicts.

Yes, if you need to merge ONNX models and prefer a CLI or simple API over writing ONNX graph code directly. Install friction is negligible, maintenance is active, and the MIT license poses no restrictions. No known vulnerabilities. Suitable for production use in model composition workflows.

## Install

pip install snc4onnx
uv add snc4onnx
poetry add snc4onnx

## Installing snc4onnx

Before you install: Low install friction with no runtime dependencies. Active maintenance as of February 2026, though repository shows modest engagement (17 stars).

License in practice: MIT License permits commercial and private use with minimal restrictions; suitable for most projects.

Quickstart:

pip install snc4onnx

from snc4onnx import combine

combined = combine(
    srcop_destop=[['output', 'flow_init']],
    op_prefixes_after_merging=['init', 'next'],
    input_onnx_file_paths=['model1.onnx', 'model2.onnx']
)

Requires ONNX models to be valid and operator names to be correctly specified; mismatched connection points will cause merge failure.

Verify before relying:

- Whether the tool handles complex multi-branch model topologies or only sequential connections.
- Performance characteristics when merging large models or many models in sequence.
- Whether intermediate validation or error recovery is available during multi-step merges.

## Package facts

- License: MIT License (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 84.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

merge onnx models, combine neural networks, onnx model fusion, join onnx graphs, network concatenation tool, onnx, model-composition, neural-networks

[View on SkillFed](https://skillfed.io/packages/snc4onnx) · [View on PyPI](https://pypi.org/project/snc4onnx/)