serverkit 0.3.3

Fluent Python SDK and shell for Linux server operations

ServerKit

A fluent Python SDK and shell for Linux server operations.

Inspect processes, logs, disk, services, and more through chainable objects and saved workflows — instead of memorizing ps, grep, and one-off shell pipelines. Run commands interactively in the REPL, automate in scripts, or ask questions in plain English with local Ollama.

Version 0.3.3 · Python 3.10+ · Linux

pip install "serverkit[rich]"

serverkit on PyPI · User guide · Examples

---

Why ServerKit

Instead of…	ServerKit gives you…
ps aux | awk … | grep …	server.processes().memory_above(500).sort_by_memory().summarize()
Ad-hoc audit scripts you rewrite every time	Named workflows in ~/.serverkit/workflows/, importable from a built-in catalog
SSH + shell one-liners per host	Server.connect("host") with the same API locally and remotely
Piping logs through grep and tail	server.logs("app.log").errors().match(r"timeout").summarize()

Filters run eagerly; .summarize(), .display(), and .all() are terminal — you always know when work is done.

---

Install

Recommended (SDK + interactive shell + formatted tables):

pip install "serverkit[rich]"

Everything (SSH remote, Docker, AI, Rich):

pip install "serverkit[all]"

Extra	Adds
[rich]	Formatted tables for .display() and the REPL
[remote]	Server.connect() / REPL connect over SSH
[docker]	Container listing, logs, and stats
[ai]	Server.ask() and REPL ask … via Ollama
[dev]	pytest and test dependencies

Core install (pip install serverkit) includes the SDK and REPL with plain-text .summarize() output. For the full interactive experience, use at least [rich].

After install, launch the shell:

serverkit

Config is created automatically at ~/.serverkit/config.json on first use (commented template — edit values or uncomment optional lines).

---

Quick start

Python SDK

from serverkit import Server

server = Server()

# Processes and memory
print(server.processes().memory_above(500).sort_by_memory().summarize())
print(server.memory().summarize())

# Logs
print(server.logs("/var/log/syslog").errors().tail(50).summarize())

# Disk and services
print(server.disk().usage_above(80).summarize())
print(server.systemctl().list_units().active().summarize())

# Workflows — import a catalog template and run it
server.import_workflow("memory_audit")
server.run("memory_audit")

# Or build and save your own
server.workflow("audit").processes().memory_above(1000).summarize().save()
server.run("audit", dry_run=True)

Remote hosts

Requires [remote]:

from serverkit import Server

with Server.connect("vm1.example", user="deploy", key_path="~/.ssh/id_ed25519") as remote:
    print(remote.processes().memory_above(200).summarize())
    print(remote.memory().summarize())
    remote.service("nginx").status()
    remote.run("memory_audit")

Interactive shell

serverkit

help
memory
processes().memory_above(500).summarize()
catalog
import memory_audit
run memory_audit
connect vm1.example --user deploy --key ~/.ssh/id_ed25519
disconnect
exit

Natural language (optional)

Requires [ai] and a running Ollama instance:

from serverkit import Server
print(Server().ask("show processes with cpu above 10 percent"))

ask list processes with cpu above 10 percent
ask largest files in /var/log limit 10

The AI layer routes requests through the SDK — it does not execute arbitrary shell from the model. Common phrases like CPU/memory thresholds use deterministic parsing so small models stay reliable.

---

What's included

• Resource facades — processes, logs, memory, disk, network, ports, systemd, cron, users, environment, Docker
• Fluent collections — chain filters, then .summarize(), .display(), or .all()
• Workflow engine — JSON workflows (schema_version: 2), catalog templates, dry-run support
• SSH remote — broad parity with local metrics and workflow execution on remote hosts
• REPL — serverkit CLI with completions, fluent chains, and connect / disconnect
• AI assistant — intent routing, diagnostics, and workflow generation via Ollama

---

Documentation

Guide	Description
User guide	Mental model, SDK, REPL, remote, AI, troubleshooting
Examples	Runnable sample scripts

---

Development

From a clone:

python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest

Integration tests (live OS): pytest -m integration

---

License

MIT — see LICENSE.