# vllm-omni (PyPI package)

> A framework for efficient model inference with omni-modality models

- Version: 0.32.0
- Released: 2026-10-28
- License: Apache-2.0 License
- Monthly downloads: 21,203
- Source repository: https://github.com/vllm-project/vllm-omni ([in the GitHits index](https://index.githits.dev/repositories/github.com/vllm-project/vllm-omni/))

## Pages

- [vllm-omni on GitHits](https://index.githits.dev/packages/pypi/vllm-omni/)
- [Direct dependencies of vllm-omni 0.32.0](https://index.githits.dev/packages/pypi/vllm-omni/0.32.0/dependencies/)
- [Security advisories of vllm-omni 0.32.0](https://index.githits.dev/packages/pypi/vllm-omni/0.32.0/security/)
- [Documentation of vllm-omni 0.32.0](https://index.githits.dev/packages/pypi/vllm-omni/0.32.0/docs/)

## Use it from a coding agent

GitHits gives coding agents version-aware access to vllm-omni: they can search, grep and read its source and documentation, and check its dependencies and security advisories, through MCP or the CLI instead of guessing from training data.

Connect an agent with one command, which detects installed agents, configures them and signs in:

```sh
npx githits@latest init
```

- [Remote MCP server](https://mcp.githits.com) (Streamable HTTP)
- [Agent documentation index](https://docs.githits.com/llms.txt)
- [API overview](https://docs.githits.com/api/overview)
- [Explore the GitHits index](https://index.githits.dev/)