Xenon
Terminal AI coding agent that separates machine-verified results from items needing human check.
Tested on Linux
Use it
Linux x86_64
You need first
- Runtimes:
python>=3.10 - Keys or accounts: LLM API key
Install Xenon from release wheel via pip
pip install https://github.com/xianyu-sheng/Xenon/releases/download/v0.9.4/xenon_harness-0.9.4-py3-none-any.whl
Check that it works
xenon --versionThe output should contain “xenon 0.9.4”.
First thing to try
xenon < /dev/nullRun xenon in your project folder to start the coding agent in the terminal.
What it does
- Record tool runs in a hash-chained evidence logNot tested
- Retry empty answers and list items to check by handNot tested
- Rewind and fork the conversation session treeNot tested
- Require approval for actions outside the workspaceNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 0 of 4 features work
- Works offline
Windows and macOS: the project says it runs there; pickfit hasn’t tried it.
Good for
An alternative to Aider, Claude Code, OpenCode and Crush.
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 0.9.4
Differences the test machine hit
- README: `xenon` starts the session directly.
In the test: Pipe EOF via `xenon < /dev/null` so the REPL sees end-of-input and exits 0 with banner plus API-key hint; a plain interactive `xenon` waits for input and had to be killed after 20s.