Early access release
Slo is a headless terminal-based coding agent that applies your team's rules, patterns and review standards on every task. Automatically. Say "dynamic workflows" and it fans the work out across a fleet of supervised sub-agents.
Available for macOS and Linux
$ slo
✦ Reading skills/INDEX.md…
you: Write a spec for the payments module
⏺ Loading context
⏺ Running PRE hook
⏺ Writing spec
✓ Saved to specifications/01-payments/
→ Ready. Ask me to build it.
Structured prompt documents covering routers, services, tests, git ops, feature specs, code review, and more. Slo loads the right ones automatically based on what you ask.
Place a SKILL.md in ~/.slo/user_skills/ and it shadows the built-in. Your conventions win without touching the core.
Use environment variables to point to your extended context. If you don't have one ask Slo to build it for you. And, let Slo maintain the context for you. Slo, refers to this context on each turn so, the more business, and domain context you provide, the better Slo gets.
New in this release
Slo now orchestrates entire workflows across parallel sub-agents, runs fully headless for scripts and CI, and fires itself on a schedule using your OS scheduler.
Say “dynamic workflows” and Slo becomes an orchestrator — decomposing a workflow into phases, dispatching each to its own focused sub-agent, and synthesizing the results.
Learn more →Headless & server agentRun slo run with no TUI. Every stdout line is a validated JSON event you can pipe through jq — perfect for CI, scripts, and long-lived server agents.
Persist a prompt plus a trigger and let launchd (macOS) or cron (Linux) run it — including recurring check-ins on a live orchestrator's sub-agents.
slo visualize — watch every sub-agent in a dynamic workflow progress live.Also new: a PDF document-generation skill — ask Slo to turn a report, brief, or portfolio into a print-ready, field-guide-style PDF.
Models
Slo works with the latest model releases, ready to power your workflows.
Currently available for macOS and Linux.
curl -fsSL https://get-slo.sh/install.sh \
| SLO_USERNAME='EarlyAccessUser' \
SLO_TOKEN='gldt-e4uTk1JUrKGdzBY_Gzqb' bashPaste the command above with your credentials. The slo binary and skill library install in seconds.
Run slo in your project directory. On first launch, connect your model provider via /auth.