Skip to main content
If you do the same AI task twice, make it a workflow. Every example on these pages is a complete one you can run right now — the whole gallery ships inside the binary:
nika examples run showcase/t1-meeting-actions --model mock/echo   # offline · zero key
The YAML you read here is projected verbatim from nika-spec/examples/showcase/, where each file passes the same conformance gate as the spec’s own test suite. If a file is on this page, it parses — and it runs. These workflows ( starters · chains · fan-outs · epics) plus the foundation set exercise all builtins, the verbs, and providers — cloud, local and mock.

How to read a workflow diagram

Tasks are colored by verb, the same palette everywhere in the Nika universe:

T1 · Starters

Four tasks or fewer, one chain or one diamond. You can run one of these in your first five minutes.

Standup digest

Engineering — your standup note, written from what you actually committed yesterday.

Meeting actions

Every office — transcript in, tracker-ready typed action items out.

Price watch

E-commerce — a price-drop alert with zero model calls.

Social repurpose

Marketing — one post becomes a thread, a LinkedIn post and a newsletter blurb, in parallel.

T2 · Chains

Four to six tasks. Structured outputs, human gates, and the data builtins start pulling their weight.

Release notes

Engineering / DevRel — git log → typed notes → CHANGELOG edited in place → team pinged.

SEO content brief

SEO / Content — a brief grounded in what your competitor actually published.

Invoice chaser

Finance / Freelance — reminders drafted, and NOTHING goes out until a human says yes.

Support triage

Customer support — the overnight queue classified, first replies drafted, urgent ones escalated.

Contract guard

Legal — clause extraction on a LOCAL model. The contract never leaves the machine.

ETL quarantine

Data engineering — bad batches degrade to a quarantine file instead of killing the pipeline.

Release radar

DevOps — dependency feeds diffed against last run. Only the NEW ships reach you.

T3 · Fan-out

for_each over collections the workflow discovers at runtime, with retry, bounded parallelism and jq fan-ins.

Competitor radar

Strategy / PMM — everything they shipped last week, read in parallel, one Monday brief.

Localization factory

Product / i18n — the whole docs tree translated, voice intact, zero copy-paste.

Config drift sentinel

SRE / Platform — only unsanctioned prod drift wakes anyone. jq decides; the model explains.

PR review fan-out

Engineering — a read-only review agent per changed file. A swarm with a leash.

Resume screener

HR / Recruiting — one local-model rubric per candidate, ranked deterministically. PII stays home.

T4 · Epic

Multi-stage pipelines: agents under budget, human gates, runs that report on themselves.

Deep research brief

Research / VC — plan → budgeted research agent → thinking-model synthesis. Auditable end to end.

Incident war room

SRE / On-call — evidence in parallel, a typed timeline, and a recovery check before the draft.

CEO Monday brief

Founders / Execs — market + engineering + KPIs synthesized — and the ping tells you what the run cost.

Release train

DevOps / Release — parallel gates, a human GO, a hold until the window. Time as a first-class citizen.

Find the example that teaches X

Looking for one construct? This index is generated from the workflow files, so it cannot go stale.
ConstructTaught by
fan-out over a collectioncompetitor-radar · localization-factory · pr-review-fanout · resume-screener
bounded concurrencycompetitor-radar · localization-factory · pr-review-fanout · resume-screener
collect-errors batchescompetitor-radar · localization-factory · pr-review-fanout · resume-screener
CEL conditional gateprice-watch · etl-quarantine · invoice-chaser · release-radar · support-triage · config-drift-sentinel · resume-screener · incident-war-room · release-train
transient-error retrycompetitor-radar · config-drift-sentinel · incident-war-room · release-train
task time-boundcompetitor-radar · release-train
cleanup that always runsceo-monday-brief · incident-war-room · release-train
fallback recoveryetl-quarantine · release-radar · competitor-radar · config-drift-sentinel · localization-factory · pr-review-fanout · resume-screener
scoped aliasingsocial-repurpose · resume-screener
jq output bindingsprice-watch · release-radar · seo-content-brief · competitor-radar · ceo-monday-brief
typed (structured) outputmeeting-actions · contract-guard · release-notes · seo-content-brief · support-triage · pr-review-fanout · resume-screener · deep-research-brief · incident-war-room
extended thinking budgetceo-monday-brief · deep-research-brief · incident-war-room
vault-backed secretsprice-watch · release-notes · support-triage · competitor-radar · config-drift-sentinel · ceo-monday-brief · incident-war-room · release-train
typed workflow inputsmeeting-actions · contract-guard · seo-content-brief · config-drift-sentinel · deep-research-brief · release-train
typed workflow outputs (callable)meeting-actions · price-watch · social-repurpose · standup-digest · contract-guard · etl-quarantine · invoice-chaser · release-notes · release-radar · seo-content-brief · support-triage · competitor-radar · config-drift-sentinel · localization-factory · pr-review-fanout · resume-screener · ceo-monday-brief · deep-research-brief · incident-war-room · release-train
default-deny tool grantspr-review-fanout · deep-research-brief
structured exec captureincident-war-room · release-train
sovereign local modelcontract-guard · resume-screener

The rules every example obeys

  • One verb per task · infer, exec, invoke or agent.
  • Every reference is an edge · ${{ tasks.X }} anywhere requires depends_on: [X]. The DAG has no invisible edges.
  • mock/echo by default · files run with zero API keys wherever the job allows; real providers only where the job demands one; local models where the data is sensitive.
  • Projection, not copy-paste · these pages and the nika.sh use-cases explorer render the same files, projected by showcase-projector.py. Any drift fails CI.

Write your first workflow

Ready to write your own? Start from the guided first-workflow walk.