tapestry-cli

The command-line interface for Tapestry — project intelligence for AI-native teams.

Wire a project into the platform, then open the live console. Stdlib-only, no third-party dependencies, no model downloads — it works the moment pip install completes, on Windows, macOS, and Linux.

Install

pipx install tapestry-cli

(or pip install tapestry-cli)

Commands

tapestry onboard      Onboard the current directory and write
                      .claude/settings.json with the Tapestry plugins enabled
                      (the recommended quickstart path).
tapestry observatory  Open the Observatory console in a browser.
tapestry init         Register the project and write .env / .mcp.json /
                      .project-intelligence/ (granular path; does not touch
                      .claude/settings.json).
tapestry version      Print version and platform info.

Quickstart

pipx install tapestry-cli
tapestry onboard
tapestry observatory

The agents pick up memory, telemetry, and the discipline rules from there. See the setup guide for the full walkthrough.

Configuration

The CLI is URL-env-driven — no hardcoded hostnames. It reads:

• LOOM_MEMORY_MCP_URL / LOOM_MEMORY_URL — the memory MCP endpoint
• LOOM_PROJECT_ID — the project identifier
• TAPESTRY_OBSERVATORY_URL — override the Observatory URL opened by tapestry observatory

Notes

tapestry is the primary entry point; loom is kept as a deprecated alias during the transition. The Python package is named tapestry_cli.

Provenance: lifted verbatim from the-loom/tapestry-cli/ during the Tapestry monorepo migration (Step 5a, 2026-06-21) — the cross-platform replacement for the old PowerShell new-loom-project.ps1 scaffolder.