memcli

A CLI for interacting with Memrise community courses

Install

npm install -g memcli
# or
bun add -g memcli

Configuration

Set your Memrise credentials via environment variables:

export MEMRISE_USERNAME="your_username"
export MEMRISE_PASSWORD="your_password"
# Optional
export MEMRISE_CLIENT_ID="your_client_id"

Alternatively, you can pass credentials explicitly with --username and --password flags.

Usage as agent skill

• add dvcrn/skills as skill repository which should discover the memrise skill
• or add the SKILL.md from this repo to your agent

Examples

List your teaching courses:

memcli courses

List words in a course:

memcli words 123456

List words in a specific level (limit to 10 items):

memcli words 123456 --level 1 --limit 10

Add a new word to a course:

memcli add-to-course 123456 --field "1=Hello" --field "2=Bonjour"

Search a pool for existing items:

memcli search-pool 7778482 --field "1=Hello"

Output Mode

The CLI defaults to a human-readable text output. For scripting, use JSON:

memcli courses --output json

Commands

• courses / my-courses – List your teaching courses
  • --limit <n>
  • --offset <n>
• course-by-id <course-id> – Fetch course details by ID
• course-by-slug <slug> – Fetch course details by slug
• course-levels <course-id> – List levels in a course
• course-columns <course-id> – Show course column configuration
• words <course-id> – List words/items
  • --level <index> – List items in a specific level (1-based index)
  • --limit <n> – Limit number of items
• learnable <learnable-id> – Fetch a single learnable item by ID
• get-pool <pool-id> – Fetch pool information
• search-pool <pool-id> – Search a pool
  • --field <key>=<value> (repeatable)
  • --exclude <learnable-id> (repeatable)
  • --original-only
• add-to-course <course-id> – Add a thing to a course (defaults to first level)
  • --field <key>=<value> (repeatable)
  • --level-index <n> (default 0)
• add-to-level <level-id> – Add a thing to a specific level ID
  • --field <key>=<value> (repeatable)

You can also pass --columns '{"1":"value"}' as a JSON string instead of repeated --field flags.