Real-world scenarios

Use Cases

How Omnus changes daily workflows for solo builders and small teams managing multiple projects.

// Scenario 01

Patch 6 repos overnight

Before

You wake up to 6 Dependabot emails about a critical Next.js CVE. You spend 4 hours over several days figuring out which repos are affected, creating branches, bumping versions, fixing test failures, forgetting repo 5.

With Omnus

Dependency audit detects CVE across 6 repos at 2am. Campaign runner patches, tests, creates PRs. 2 auto-deployed by morning, 3 PRs ready for review, 1 flagged for manual migration. 12 minutes of your attention.

Campaign: next@14.2.29 CVE-2026-3891
omnusdeployed
cullr-appdeployed
terso-sitePR ready
istqb-appPR ready
kindlmann-sitePR ready
woeker-apimanual
6/6 complete — 2 deployed, 3 awaiting review, 1 manual

Telegram

Voice note0:10

"Switch cullr to edge runtime for the matching endpoint, the cold starts are killing conversion"

Transcribed, classified, routed to cullr-app
~/projects/cullr
$ terso sync
  CURRENT_CONTEXT.md  fresh
  RECENT_DECISIONS.md  1 new

  Agent sees: "Switch to edge runtime
  for matching endpoint (cold start issue)"

// Scenario 02

Voice note to agent context

Before

You have an idea while walking. You try to remember it until you are at your desk. Half the context is lost. When you open Cursor, you re-explain everything from scratch.

With Omnus

Hold the Telegram voice button, speak for 10 seconds. Omnus transcribes, classifies, routes to the right project. Run terso sync — your agent already knows.

// Scenario 03

Portfolio health at a glance

Before

You have 30 deployed sites. You check them manually when you remember. SSL expires, dependencies go stale, Lighthouse scores drop — you find out from users, not monitoring.

With Omnus

Nightly health checks across all sites. Morning briefing via Telegram: "30/30 healthy, 1 SSL expiring in 26 days, Lighthouse dropped 14 points on kindlmann.com." One tap to approve auto-fix.

Morning briefing — Mar 15
30/30 sites healthy
SSL expiringkindlmann.com — 26 days
Lighthouse dropkindlmann.com — 94 to 78
Deps outdated2 repos — minor only
DNS verifiedall records match
Next check: tonight 02:00 UTC
~/projects/cullr
$ terso sync
  Synced 5 files to .terso/generated/

$ ls .terso/generated/
CURRENT_CONTEXT.mdStack, architecture, current status
RECENT_DECISIONS.mdWhy we chose Supabase Auth, edge runtime
SHARED_OPS.mdDeploy to Cloudflare, coding standards
KNOWN_RISKS.mdOpen issues, recent incidents
context.jsonMachine-readable metadata for agents
Cursor / Claude Code reads these automatically

// Scenario 04

New developer, zero onboarding

Before

Freelancer joins for 2 weeks. You spend 3 hours explaining the project, the architecture, the deployment process, the coding standards, the recent decisions.

With Omnus

"Clone the repo, run terso sync." Five files appear. Cursor reads them. First question on day 1 instead of 5+ interruptions over 3 days.

Stop re-explaining yourself to AI agents.

Capture once, use everywhere.