A coding agent in your terminal.
One binary, no runtime, no browser. Free and open source.
What it is
The same agent core as the desktop app, compiled to a single 6 MB Rust binary with a keyboard-first terminal interface. It shows the plan as it forms, the thinking stream, the diffs before they apply, and running token accounting so you can see what a turn is costing while it happens rather than afterwards.
Why a binary rather than a Node package
No runtime to install, no dependency tree to resolve, no version conflict with whatever else is on the machine. It matters most in the environments where this kind of tool is most useful: a server you have SSH access to, a container, a machine you do not administer.
Free and open source
The CLI is free with no subscription. The paid product is the desktop app, which adds the diagram canvas, the dependency audit, the observability surface and the command palette. If the terminal is where you work, you never need to pay anything.
Same models, same rules
Any OpenAI-compatible endpoint, local Ollama included, so the offline story is identical. Configuration is shared with the desktop app, so using both means one set of settings.
CLI and desktop app
| CLI | Desktop app | |
|---|---|---|
| Price | Free, open source | $3/month |
| Coding agent | Yes | Yes |
| Reviewable diffs | Yes | Yes |
| Diagram canvas | No | Yes |
| Dependency audit | No | Yes |
| Grafana and Prometheus | No | Yes |
| Command palette | No | Yes |
| Runs over SSH | Yes | No |
| Size | 6 MB binary | Native app |
Install
Questions
Is the CLI really free?
Free and open source, with no feature gating against the desktop app's coding capability. The desktop app charges for the visual surfaces.
Does it work over SSH?
Yes, which is one of the main reasons it exists. A single static binary with no runtime is straightforward to get onto a remote machine.
Can it use local models?
Yes, identically to the desktop app. Same configuration file.
Related
- AI Assistant for ObservabilityWorth being precise about what this is: it drives the observability tools you already run. It is not an LLM tracing platform, and if that is what you searched for, the tools below are the right ones.
- AI Assistant for GrafanaThe value is not writing the query. It is going from the query result to the line of code without leaving the session.
- PromQL Query AssistantPromQL is not difficult, it is just impossible to remember the exact shape of a rate over a histogram at the moment you need it.
- Loki Log Analysis with AIFinding the log line is half the job. The other half is working out what emitted it and why.
Last updated