What the connector does, and what stays in your AI
Available on Solo, Pro and Team.
The connector lets your AI use pmmAgent.ai directly. You ask for a battle card in a chat, and your AI runs the tool and writes the asset without you copying anything.
Available on Solo, Pro and Team.
What pmmAgent.ai does
When your AI runs a tool, pmmAgent.ai:
- Takes the content type and the brief your AI sends (up to 20,000 characters).
- Adds your Brand Box, and for a named project its PRD and pinned outputs.
- Returns the briefing: a structured brief for that tool, built to be written up.
- Counts the run against your plan, exactly as on the web.
- Ends the reply with the workspace used, which Brand Box parts applied, and how many runs you have left.
pmmAgent.ai sees only what your AI sends in the request. It does not read your other chats or files.
What stays in your AI
Your AI writes the finished asset. The document, the slide outline, the email or the post is made in your Claude or other AI session, in the format you ask for, and it stays there. pmmAgent.ai never receives the finished asset unless you ask your AI to pin it to a project.
This is also why you can ask for formats pmmAgent.ai does not produce itself: a file, a deck, a table, a message to paste into Slack.
Connector or web app
Same tools, same runs, same Brand Box. The connector saves the copy and paste. The web app suits project work on Canvas and context-heavy requests in Strategic. See Web app or your own AI: which to use when.
Pro tools need the Pro plan (or Team). On Solo, asking for one returns a message that the tool is part of Pro.
What you need
- A Solo, Pro or Team plan. On Free, the connector declines and tells you the plan is web-app only.
- An AI that supports remote connectors. Claude (web and desktop), Claude Code, Codex and Cursor are covered in these guides.