Skip to main content
A Play is a markdown instruction document stored in Bridge’s Plays Library. When you load a Play before running a command, Bridge’s AI reads those instructions first and uses them to guide how it approaches the task. Plays encode your team’s best practices and preferred workflows so the AI doesn’t have to guess — and you don’t have to re-explain how things should be done every time.

​
What a Play contains

Each Play is a plain markdown file with step-by-step instructions written for the AI. Think of it like a standard operating procedure that the AI reads and follows before it starts working. A Play can specify which fields to populate, how to name things, what to skip, and how to handle edge cases.
Plays are read-only instructions loaded into the AI’s context. They don’t execute code or call your tools directly — they shape how the AI calls them.

​
Available plays

Bridge ships with plays for the most common workflows. Below are three examples to give you a sense of what’s available.

transcript-to-tasks

Turns a meeting transcript into a structured list of ClickUp tasks. The Play instructs Bridge on how to identify action items, extract owners, set due dates from context clues, and assign tasks to the right list. Paste a transcript and Bridge does the rest.
Walks Bridge through a standard client onboarding sequence: create ClickUp tasks for each onboarding step, send a welcome email, and log the new client in your CRM. Variable inputs include the client name, contact email, and target start date.
Summarizes team progress from Slack standup messages. The Play tells Bridge how to structure the summary — what was done, what’s blocked, what’s coming next — and how to post or save the result.
To see all plays currently available in your account, type List the available plays in the Command Center or use the bridge_list_plays MCP tool if you’re connected via MCP.

​
How to use a Play

You load a Play by mentioning it in your command. Bridge recognizes the intent, fetches the Play’s instructions from the library, and applies them before executing.
1

Name the Play in your command

Reference the Play by name in the Command Center. You can combine it with your actual input in the same message.
example commands
2

Bridge fetches the Play

Bridge retrieves the Play’s instruction document from the library and loads it into the AI’s context for this command.
3

Bridge executes with the Play's guidance

The AI follows the Play’s instructions as it calls your connected tools — filling fields the right way, naming things consistently, and handling edge cases as the Play specifies.

​
Why Plays matter

Without a Play, Bridge’s AI makes reasonable guesses about how to structure its output — what to name a task, which fields to fill, how to handle ambiguity. Those guesses are often fine, but they won’t always match how your team works. Plays eliminate that variability. Once a Play exists for a workflow, every run of that workflow produces output structured the same way — regardless of who runs it or how they phrase the command.

Consistency

Every run of a Play-guided command follows the same structure and naming conventions your team agreed on.

Institutional knowledge

The Play captures how your team does things — edge cases, field mappings, preferences — so new team members get the same results as experienced ones.

​
MCP tools for Plays

If you’re using Bridge connected to Claude or Gemini via MCP, two tools give you direct access to the Plays Library from within your AI client.

bridge_list_plays

Lists all available plays in the library. Use this to discover what’s available before starting a task.Returns: play names, descriptions, types, and associated connectors.
usage in Claude or Gemini
Fetches a specific Play by name and loads its full instruction content into the AI’s context. Supports exact match and case-insensitive fallback.Input: play_name — the name of the Play (e.g., transcript-to-tasks)
usage in Claude or Gemini
These MCP tools are available when Bridge is connected to your AI client as an MCP server. In the Bridge Command Center, you load Plays by mentioning them in plain English — no tool names required.

Command Center

Where you run Play-guided commands in plain English.

Missions

Save successful Play-guided runs as reusable workflows.