Introduction

Bring Ancher into your workflow — from a coding agent or from your terminal.

Ancher is your personal knowledge base and research assistant. These docs cover the two ways to reach it from a developer workflow: Agent Skills, which give a coding agent direct access to your knowledge base, and the CLI, which puts Ancher in your terminal.

They are separate installs, but not always independent: the recommended Agent Skills path uses the CLI as the agent’s backend, while the MCP install stands on its own. The Quickstart walks through each in a few commands.

Agent Skills

Give your coding agent the skills to work with your Ancher knowledge base.

bash

npx -y skills add streamify-one/ancher-agent-skills

Works with Claude Code, Codex, Cursor, and 70 more agents. Pair it with the CLI below, which is how the agent reaches Ancher. On Claude Code, Codex, or Cursor you can instead run npx -y @ancher-ai/agent-skills install to configure the hosted MCP server and skip the CLI. Start at Agent Skills installation.

CLI

A terminal-friendly interface for saving, searching, organizing, and creating from your knowledge base.

bash

npm install -g @ancher-ai/cli
ancher login

Commands that return a record or a page print JSON; the ones that don’t — action confirmations, raw content, streamed chat prose, api passthrough — are spelled out per kind under Output. All of them authenticate from an environment variable and reject an unknown flag before a request leaves your machine, so one interface serves a person at a prompt, a CI job, and an AI agent equally well. Start at CLI installation.

Which one do I want?

If you want to	Use
Have your coding agent save and search Ancher content	Agent Skills
Work with Ancher yourself, from a terminal	CLI
Script Ancher in CI	CLI, with `ANCHER_API_TOKEN`

The recommended Agent Skills install uses the CLI as the agent’s backend. The MCP install is the way to run Agent Skills without the CLI at all, because the agent talks to the hosted server rather than shelling out.