Demos and features
This is the full catalog, grouped by the question a new developer is likely asking. If you only have a few minutes, use the curated Proof path first.
Every page below is generated from the same feature catalog used by in-app tours and QA scenarios. Recorded demos replay from cassettes; synthetic demos use no-LLM fixtures. The published site render does not improvise with a live model.
Runtime control
Agent action transcripts
Every agent call's full tool-use stream — captured once, replayed byte-identically.
Operator questions, forwarded
Agent questions become blocking operator modals — never silent defaults.
Decision-first trace detail
Verdicts, confidence, evidence, and replay — on a frozen trace.
Trace introspection, live
View modes, waterfall, decision-first detail, annotation, replay.
Try it in an existing repo
The stranger install — download, run, onboarded
A released binary, dropped into a foreign repo, profiles it and installs the toolkit without a checkout.
kitsoki doctor — provider, toolkit, and MCP readiness at a glance
One command reports whether your provider, agent toolkit, and MCP registration are actually ready.
An actionable message on a no-provider first run
No Claude CLI and no API key? kitsoki says exactly what's missing instead of silently faking it.
A first screen that only offers actions that work
The get-started screen advertises exactly the intents the running story can resolve.
The harness picker — pluggable operators, live
Switch provider, model, and reasoning effort from the run header; every call is stamped with the operator that answered it.
The kitsoki web UI, in five minutes
A story-agnostic guided walkthrough of every core surface.
Author workflows safely
Meta-mode — the machine edits its own story
Ask the running FSM about its own graph, then have it rewrite and hot-reload its YAML — live.
The story editor
Rooms, hooks, domain model, and the Agent Workbench — one inspector.
Dynamic workflows
Turn a proposal-shaped goal into a validated draft, a live session, and an exportable starter base.
Punch-list story
Run YAML worklists through Kitsoki stories with model policy and independent verification.
Goal-seeker — the loop that drives the plan through kitsoki itself
A kitsoki story that evaluates a decomposed plan, dispatches workers, integrates green work, and loops until done.
Inspect and operate runs
Live agent streaming in chat
Watch the agent think and act, streamed live into the conversation.
The async inbox
Background work notifies you when it lands — toast, badge, jump back.
Session Media Workbench
Pin the artifact, keep the conversation, move the devtools.
The state diagram, four ways
Metro stepper, ego-graph, path & horizon, and the full machine.
Video review and feedback
Chaptered playback with frame-anchored feedback flags.
The complete product tour
One continuous walkthrough — eight independently-releasable sections.
The full product walk
Catalogue to live session to reload-survival — end to end.
Real workflow case studies
Bug triage to autonomous fix
Triage a ticket, hand it to the pipeline, watch it land.
Fix a real bug in slidey, end to end
kitsoki's bug-fix pipeline reproduces, fixes, tests and validates a real slidey bug — a grid-cards narration-timing drift.
Open the PR for the slidey fix
kitsoki's PR-refinement pipeline opens, monitors and merges a pull request for the real slidey grid-cards narration-drift fix.
PM idea → PRD on the slidey repo
A product manager talks a real slidey feature — a `--notes` speaker-notes export — from a one-line idea into a published PRD, no LLM in the loop.
Architect → design on the slidey repo
An architect picks up the published slidey speaker-notes PRD and authors the design — epic, slices, ADRs, and a mermaid data-flow — all conversation-driven, no LLM.
Design → validated work plan on the slidey repo
The published slidey speaker-notes design is decomposed into three dependency-ordered, gate-bearing implementation briefs — decompose → lint → a validated work plan — no LLM.
Design pipeline walkthrough
From idea to published design doc — one conversation.
Mockup Video Studio
Produce, review, and refine a rendered video — room by room.
Weather & Climate (Starlark host)
Deterministic HTTP workflows in a sandboxed Starlark hook.