Ask in the chat. PitCrew runs the agent under your firm's rules and brings the result back to you.
You need a PitCrew account. Install the PitCrew plugin from our GitHub repository and sign in once. The plugin includes the PitCrew connector.


pitcrew-agents/pitcrew-plugin


/plugin marketplace add pitcrew-agents/pitcrew-plugin
/plugin install pitcrew@pitcrew-plugin/mcp and sign in to PitCrew MCP with your PitCrew account.Pick an example to see it play out.
Install the plugin once. It adds the PitCrew connector, so Claude can act in your PitCrew workspace, and an agent builder that works in the chat.
The actions Claude takes in your PitCrew workspace, under your PitCrew sign-in.
list_agents, start_agent_runopen_agent_input_formget_run_statuscreate_collection, upload_documentssubmit_validation_async, get_validation_resultsget_relevant_citations, get_rules_and_variablessend_email9 skills and 6 helper agents that plan, build, test and publish an agent from your description, into your Agent Crew. You approve the plan before anything is built.
Claude starts the run and shows you the result. Your firm's rules, approvals and records stay where they are.
In Claude, in plain words. No new app to open.
Back in the same chat, with a link to the full run in PitCrew.
In Claude, go to Customize, Plugins, + Add, Add marketplace, Add from a repository. Paste this and select Sync.
pitcrew-agents/pitcrew-pluginNo PitCrew account yet? Request a call
Yes. You sign in with your PitCrew account when you connect PitCrew MCP in Claude. Claude can only reach what that account can.
Yes, by default. Claude asks for your approval each time it uses a PitCrew tool. You can change that for each tool in Claude, under Customize, then Connectors.
The repository is PitCrew's official GitHub repository. It is public and MIT-licensed, so you can read every file before you add it.
No. Claude starts the run and shows you the result. The agent runs in PitCrew, under your firm's rules, and the run is kept in PitCrew with its evidence.
With the plugin, describe a new agent and PitCrew builds it. With the connector, run your published agents with inputs from the chat or a Collection, add documents to a Collection, check documents against your firm's policies, find the rule that applies, check a run's status and result, and send the result by email through the Gmail account connected in PitCrew.
The connector, PitCrew MCP, lets Claude run your published agents, add documents to Collections, check runs and read results. The plugin adds the agent builder, so you can describe a new agent in the chat and PitCrew builds it. Installing the plugin also adds the connector.
Not yet. Both are coming soon.
Describe your biggest workflow challenge. We'll show you how we'd approach it.
We'd like to understand your specific setup. Enter your details and we'll reach out.