The Shipchart CLI is live. It's a single Go binary that gives you full access to your portfolio, projects, revenue, roadmap, goals, and AI tools from the terminal. Install it, authenticate, and you're running.
brew install shipchart/tap/shipchart
shipchart auth login
shipchart dashboard
Three commands and you've got your morning briefing without opening a browser.
Why a CLI
Shipchart already has a web app, an iOS app, a REST API, and an MCP server. So why build a CLI on top of all that?
Because some things are just faster in the terminal. Checking your MRR shouldn't require opening a browser, waiting for a page to load, and clicking through tabs. shipchart dashboard takes two seconds and shows you everything: total MRR, revenue this month, active projects, stale projects, goal progress. Done.
But the real reason is agents. If you're using Claude Code, Codex, or any terminal-based coding agent, the CLI becomes a bridge between your agent and your project data. More on that below.
What it can do
The CLI covers the full Shipchart feature set. Everything you can do in the web app, you can do from the terminal.
Portfolio and projects. shipchart dashboard for the portfolio overview. shipchart projects list to see all your projects. shipchart projects show worda for the full detail on a specific project, including tech stack, integrations, recent check-ins, and roadmap.
Revenue. shipchart revenue list worda to see revenue entries. shipchart revenue add worda 450 --period 2026-03 --type mrr to log an entry. Quick, scriptable, no form to fill in.
Roadmap. shipchart roadmap list shows your Now/Next/Later board across all projects. shipchart roadmap add worda --title "Stripe webhook improvements" --horizon now to add an item. shipchart roadmap summary for a text overview.
Check-ins. shipchart checkin add worda --mood good --body "Shipped MCP OAuth, three bugs fixed" logs a check-in. shipchart checkin generate worda uses AI to generate a check-in from your recent development activity.
Goals, expenses, analytics, forecasting. All there. shipchart goals list, shipchart expenses add, shipchart analytics trends, shipchart forecast. The full toolkit.
AI tools. shipchart ai think worda runs Think With Me on a project. shipchart ai challenge <roadmap-item> gets structured critique on your feature plan. shipchart ai generate <roadmap-item> creates work items from a description.
Delegation. This is a big one. shipchart delegate run worda --runner codex --title "Add rate limiting to API endpoints" delegates a coding task to an AI agent directly from the terminal. shipchart delegate list worda shows active delegations and their status. shipchart delegate show <id> gives you the detail, including PR links when the agent opens them.
Three output formats
Every command supports three output formats via the --format flag:
Text is the default. Human-readable, formatted for the terminal. What you see when you run a command normally.
JSON (--json) returns structured data for scripts and automations. Pipe it into jq, feed it to another tool, use it in a bash script. Standard machine-readable output.
LLM (--llm) is the interesting one. It returns compact markdown optimised for AI context windows. Less verbose than JSON, more structured than plain text, and significantly fewer tokens. When you're piping project data into a coding agent, this is the format you want.
shipchart context export worda --llm | pbcopy
That gives you your full project context (tech stack, architecture, conventions, roadmap, recent activity) as compact markdown, copied to your clipboard and ready to paste into Claude, ChatGPT, or whatever you're using.
CLI as a tool for coding agents
This is where the CLI gets genuinely interesting.
MCP is great for giving AI tools native access to your data. But sometimes a CLI is actually the better choice. It's faster for simple lookups (no tool discovery overhead), it uses fewer tokens (especially with --llm output), and it works with any agent that can execute shell commands, which is basically all of them.
A coding agent like Claude Code can run shipchart roadmap list --project worda --json and instantly know what you're working on, without needing an MCP connection configured. It can run shipchart context export worda --llm and get your full project context in a format designed to be token-efficient. It can check shipchart errors show worda to see what's broken before it starts fixing things.
And because the CLI supports write operations, agents can close the loop. An agent finishes a task, runs shipchart checkin add worda --mood good --body "Completed rate limiting implementation", and the check-in is logged. It could run shipchart delegate run worda --runner codex --title "Write tests for rate limiting" and kick off the next task.
This opens up some interesting automation possibilities. A script that reviews your roadmap, picks the top item in Now, delegates it to a coding agent, waits for the PR, and logs a check-in when it's done. A morning routine script that runs shipchart dashboard, checks for errors, and surfaces anything that needs attention. Delegation chains where one agent finishes a task and queues the next one.
The CLI makes Shipchart composable. It fits into whatever workflow you're building, whether that's a simple alias or a complex automation pipeline.
Installation
Homebrew (macOS and Linux):
brew install shipchart/tap/shipchart
Go install:
go install github.com/shipchart/cli@latest
Direct download: prebuilt binaries for macOS, Linux, and Windows are available on GitHub Releases.
Authentication
The easiest way is the OAuth Device Flow, the same pattern used by gh:
shipchart auth login
This opens your browser, you sign in to Shipchart, enter a code, and you're authenticated. Token gets stored in ~/.shipchart/config.json. No manual token management.
For CI/CD or scripts, you can pass a token directly:
shipchart auth login --token
Or set the SHIPCHART_TOKEN environment variable.
Project-aware by default
Drop a .shipchart.yml file in your repo root and the CLI automatically scopes commands to that project. Run shipchart roadmap list inside the Worda repo and it shows Worda's roadmap without needing a --project flag. Same pattern as .github/ or .vercel/.
MCP or CLI? Both.
The MCP server and the CLI are complementary. MCP is best when you want your AI tool to have always-on, native access to your data inside conversations. The CLI is best for quick lookups, scripting, automation, and when you want to keep token usage lean.
Use MCP when you're having a conversation with Claude and want it to naturally pull in project data. Use the CLI when you're writing a script, building an automation, or want your coding agent to grab specific data without the overhead of a full MCP connection.
Or use both. They hit the same API, return the same data, and work together without conflict.
Full docs at shipchart.app/docs/cli. Free tier users get full CLI access. Install it, run shipchart dashboard, and see your whole portfolio in two seconds.