Mattermost CLI (mm)
Read and search Mattermost from the command line. JSON output by default, --human for markdown.
Setup
Check if mm is already available:
mm whoamiIf that works, skip to "Start here" below.
If mm is not found, install it. Pick whichever works in your environment:
pip install mattermost-cli # adds `mm` to PATH
# or
pipx install mattermost-cli # isolated global install
# or
uvx --from mattermost-cli mm # run without installing (needs uvx)Then authenticate (one-time):
mm login --url https://your.mattermost.server
mm whoami # verifyThe login command will prompt for credentials interactively. If your server supports Personal Access Tokens (Profile > Security > Personal Access Tokens in the Mattermost UI), you can skip the prompt:
mm login --url https://your.mattermost.server --token YOUR_TOKENFor environment variables, multiple servers, and troubleshooting, see references/setup.md.
Start here: mm overview
Always run this first. It returns mentions, unread channels, and active channels in a single call.
mm overview # last 6 hours (default)
mm overview --since 1d # last 24 hoursThe response has three sections:
- mentions - posts that @-mention you, with root message context when it's a reply
- unread - channels with unread messages, sorted by count
- active_channels - channels with recent posts, sorted by recency
Each entry includes a ref field you can pass directly to other commands.
Reading messages
mm messages <channel> # last 30 messages, chronological
mm messages <channel> --since 2h # messages from last 2 hours
mm messages <channel> --threads # thread index: root + reply count + last reply
mm messages @username # DMs with someone<channel> accepts a name (off-topic), @username for DMs, or a channel ID (for group DMs from overview output).
Threads
Every post includes a thread_id. Use it to read the full conversation:
mm thread <thread_id> # root + last 9 replies
mm thread <thread_id> --limit 0 # entire thread
mm thread <thread_id> --since 1h # just recent replies (root always included)Searching and mentions
mm search "deployment issue"
mm search "from:alice in:devops after:2025-01-01"
mm mentions # @-mentions in last 24h
mm mentions --since 3dMentions for replies include a root field with the original message, so you know what "this" or "it" refers to without a follow-up call.
Channel context
mm channel <name> # purpose, header, member/pinned count
mm pinned <channel> # important/pinned posts
mm members <channel> # who's here + online status
mm channels --since 6h # all channels with recent activity
mm channels --type dm # just DMsPeople
mm user @someone # profile, role, status, timezoneKey JSON fields
Every post includes these fields so you can navigate without guesswork:
| Field | What it's for |
|---|---|
thread_id | Pass to mm thread to read full conversation |
ref | On channel entries; pass to mm messages |
is_bot / bot_name | Webhook and bot posts are flagged automatically |
root | On reply-mentions; the original message being replied to |
is_reply / reply_count | Thread structure |
reactions | Emoji counts like {"+1": 3, "white_check_mark": 1} |
Bot posts from webhooks automatically extract alert content from Slack-format attachments, so you see the actual alert text instead of empty messages.
Further reading
- references/scenarios.md - real use cases: "what did I miss?", "summarize this channel", "is this resolved?", "find that thing someone said"
- references/workflows.md - command sequences: morning triage, incident investigation, channel discovery
- references/commands.md - full command reference with all options and flags