Installation

Crew ships as a Python package that bundles the React dashboard. There are four ways to install it; pick whichever fits how you'll run it. All builds are driven by plain pip + npm/Vite + pytest. No proprietary tooling.

Crew drives an LLM through kiro-cli over the Agent Client Protocol. It is the only provider (agent.provider = acp). The dashboard guides Kiro CLI setup and sign-in before chat opens. kirocrew doctor reports whether the binary is found and logged in.

If the Gateway detects a Kiro CLI without the required acp command, setup stops at Kiro CLI update needed. Select Update Kiro CLI to run kiro-cli update on the Gateway host, then retry the check.

Semantic memory needs no setup. Embeddings run in-process and the Gateway downloads its embedding model (~610 MB) in the background on first start, verifies it, and stores it under ~/.kiro/crew/models. Until the model lands, memory search falls back to keyword matching and picks up embeddings automatically without a restart. Set KIROCREW_EMBED_MODEL_URL to point at a mirror for airgapped installs.

Crew supports Windows natively. The signed installer is the simplest path; a source checkout can use CPython 3.12, a virtual environment, and pip install -e . tzdata.

py -3.12 -m venv .venv
.venv\Scripts\activate
pip install -e . tzdata
python -m kiro_crew gateway

The Windows Gateway cold start is noticeably faster in 0.7. Computer Use works through UI Automation, and isolated pods run through Task Scheduler when the current user can create a scheduled task. See the CLI reference for pod commands; pod api is the only pod verb unavailable on Windows because its private request channel depends on a Unix-domain socket. Use pod token and the printed URL with your own HTTP client instead.

The signed installer updates in place and relaunches the new build. The interactive web terminal and sign-in PTY use their platform-specific implementation or report a clear unsupported state rather than failing silently.

Desktop and managed installations surface updates in Settings → About, the update dialog, and What's New. When host approval is required, each surface shows the same command and the time remaining in the approval window.

The Gateway checks for source updates at startup and every twelve hours. Automatic apply first pauses new turn admission, then waits for active turns, scheduled runs, subagents, Task Runner steps, and workflows to become idle. It applies only after that work settles; an active workload defers the update rather than being cancelled.

Before moving the checkout, Crew verifies that the target revision supports the Python interpreter in the current virtual environment. A mismatch is refused with the environment and interpreter path named — a Python below the 3.12 floor is rejected outright. A dirty or diverged checkout is also refused rather than reset.

Run an update manually when needed:

kirocrew update

For wheel and cli.sh installs, the updater installs into a fresh, signature-verified tree and switches atomically. Docker images remain immutable; pull a newer tag and recreate the container as described in Running 24/7.

When Crew changes a default value between releases, your stored config may still hold the older default. Run kirocrew config defaults to list stored values that still hold superseded defaults. Use --adopt to take the current defaults or --keep to record the stored values as intentional.

This applies after upgrading, particularly if the voice stt_model or sandbox settings were set by an earlier version.

For the one-line wheel, source build, and Docker installs, set up and verify after installing:

kirocrew setup            # interactive wizard: data dir, agent, credentials
kirocrew doctor           # verify everything is wired up
kirocrew gateway          # start the server → http://localhost:5476

kirocrew doctor reports the status of the kiro-cli binary, its authentication, Ollama/embeddings, Slack tokens, config validity, and MCP servers.

All persistent state lives under ~/.kiro/crew/ (override with KIROCREW_HOME):

~/.kiro/crew/
├── config.json          # user configuration
├── .env                 # Slack tokens, owner ID, other credentials
├── workspace/
│   ├── memory/          # preferences.md, projects.md, history/
│   ├── lessons.jsonl    # learned corrections
│   └── knowledge/       # ingested documents (FTS5 + vectors)
├── conversations/       # JSONL session logs
├── crons.json           # scheduled jobs
├── audit.log            # bash command audit trail
└── agents/              # generated kiro-cli agent configs

Both KIROCREW_HOME and KIROCREW_PORT are validated at CLI entry, not stored as config keys. Pass --port on the CLI to override the env var per-invocation.

Install with pip install "kirocrew[<extra>]":

Piper TTS is the default text-to-speech provider and runs locally — no extra install needed.