InferCrane
Evidence-gated system to deploy and manage vLLM/SGLang inference behind one OpenAI-compatible endpoint.
Tested on LinuxKnown vulnerabilities
Use it
Before you install: known vulnerabilities
pickfit found 6 known high or critical vulnerabilities in the packages a fresh install brings (OSV.dev, 3 Nov 2026). Weigh this before you install it.
go:github.com/go-jose/go-jose/v3 v3.0.5Go JOSE: DoS when parsing JWKs that contain RSA keys (GHSA-mgcc-cww5-fp2q, high, no fix yet)npm:undici 7.29.1undici vulnerable to TLS certificate validation bypass via dropped tls options in Agent (GHSA-2m5v-pch6-74hx, high, fixed in 7.30.1)
See 4 more
npm:undici 7.29.1— GHSA-4w7w-5xrc-9h26 (high), fixed in 7.30.1npm:undici 7.29.1— GHSA-qrxm-fqhp-jwvm (high), fixed in 7.30.1npm:undici 7.29.1— GHSA-xjj6-f33q-929c (high), fixed in 7.30.1npm:zod 4.6.5— GHSA-mrhx-c7xc-m3w5 (high), fixed in 4.6.6
Linux x86_64
download and inspect CLI archive
curl -sL -o /tmp/infercrane.tar.gz 'https://github.com/infercrane/infercrane/releases/download/v1.0.0-rc.1/infercrane_1.0.0-rc.1_linux_amd64.tar.gz' && mkdir -p ~/.local/bin && tar -xzf /tmp/infercrane.tar.gz -C /tmp/install CLI binary and check version
cp /tmp/infercrane ~/.local/bin/infercrane && chmod +x ~/.local/bin/infercrane && infercrane --help
Where the README and reality differ
- README: npm install '@infercrane/sdk@1.0.0-rc.1' and pip install infercrane wheel install the CLI
What worked: Linux CLI binary from release archive infercrane_1.0.0-rc.1_linux_amd64.tar.gz extracted and copied to ~/.local/bin/infercrane
Check that it works
infercrane --versionThe output should contain “infercrane version 1.0.0-rc.1”.
First thing to try
infercrane workload init ./support --recipe qwen3-8bRun infercrane --help to see the available commands, or start a workload with infercrane workload init <directory> --recipe <recipe>.
What it does
- Optimize existing inference workloadsNot tested
- Adopt existing inference endpointsNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 0 of 2 features work
- Works offline
macOS: the project says it runs there; pickfit hasn’t tried it.
Good for
- deploy open-weight models into your cloud or Kubernetes
- optimize workload performance and cost on exact hardware
- run agent tasks in persistent sandbox
- trace production inference failures and latency
An alternative to Raw vLLM and Kubernetes.
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 30 Oct 2026
- Version 1.0.0-rc.1