Agents run proposals, onboarding, billing checks and reconciliations under your firm's rules. A person approves the key steps. Every run is kept.
Create, edit and manage your crew's agents, with pit stops and actions.
What you need for this run
Run 4812-ferris · Oct 12, 9:12 AM · 0m 00s · Nina Brooks
Every run, build, evaluation and output, in one place.
Fidelity
Google Drive
Gmail
Slack
Start a run, check its status, approve a step, or ask a question on the data from whatever is already open. PitCrew runs an MCP server for AI assistants, an app in Slack, a mailbox for email, and an API for your own systems. Every door leads to the same run.
Same run, same rules, same record. The request can come from a chat, a channel, an inbox or your own system. The policy checks, the approval and the run history are the same.
Your AI assistant can use PitCrew directly. Add PitCrew to Claude in about two minutes. Ask it to build an agent or start a run, and the result comes back in the chat.
Add PitCrew to Claude →mcp.gopitcrew.comThis page also works with browser AI agents. They can book a call with us or browse our agents directly.
Proposals, follow-ups, client requests, onboarding packets, fee validation, reconciliations, drift checks, report packs, across the front, middle and back office. Pick the ones your firm needs; a PitCrew expert configures each to your rules, systems and approval path.
Open any agent and see how it works. What it does, what it needs, what it gives you, and every pit stop.
Most runs start on their own. A file landing in a folder, a schedule, a new record in the CRM or an email with an attachment can start a run, so nobody has to remember to press Run. One run can cover every account in the book, not a sample, and when it finishes the next agent can pick up its output.
See the agent libraryCreate, edit and manage your crew's agents, with pit stops and actions.
Describe it in Claude or in the PitCrew app. PitCrew builds it, tests it and adds it to your Agent Crew.
Most firms start with us building the first agents. Then operations, service and advisory teams build the next ones themselves, and the same rules and run history apply from the first run.
PitCrew fills in what it already knows from your CRM. The client completes the rest from one private link, and your team sees where every send stands.
Use it for account opening and transfer forms, onboarding and KYC, proposals, and lead capture from your website.
94 fields · 9 screens · Updated Oct 28
Held at Charles Schwab & Co., Inc. Managed by Nina Brooks, Larkspur Wealth.
Connect your systems or upload files. PitCrew matches them once, and every agent reads the same record.
CustodianNo re-matching on the next run.
Conflicts are flagged, never guessed.
Access follows your roles.
Everything your agents can use: connected apps, built-in capabilities and your firm's own APIs.
Fidelity
Slack
Google Drive

Custodians, CRMs, portfolio platforms, document stores, email and calendars, through standard APIs. Nothing gets replaced. Add your firm's own API in the Garage and every agent can use it.
40+ integrations, 127 built-in capabilities. Reading PDFs, working with sheets, researching the web, running the numbers. Every agent has them from day one.
Upload the billing manual, the compliance manual, the trading procedures. PitCrew turns each one into named rules with an owner and a version. Agents check their work against those rules and tell you which rule decided what.
Three layers. Your firm's policies, each client's signed documents, and the regulations that apply to you. Policy checks use formal verification (AWS Automated Reasoning), so a rule either holds or it doesn't.
Firm policies, client documents and regulations, turned into rules every agent checks against.
"A proposal shall flag any asset class whose allocation differs from the selected model by more than 5 percentage points, and state the rationale for each."
Exam prep used to cost $80-100K and eat a full week before every SEC examination. With PitCrew, the evidence assembles itself. Every agent run produces documented output: analysis reports, exception logs, decision rationale, source citations. These artifacts land in your evidence vault organized by category, time period, and regulation.
When your next exam arrives, you're not scrambling to prove the work was done. It was documented as it happened. Share with your examiner when every item is complete, not a moment before.
Assembled from 1,412 runs between Apr 1 and Jun 30 · Updated Jul 2
Kitt sits on every run, every collection and the firm as a whole. Ask why an account was flagged, whether an agreement is fully executed, or which households are outside their band. The answer cites the run, the page and the rule.
A slash command starts an agent from the same box. Type /run and the agent name, and the run is kept like any other.
Two households, both from last night's run:
A PitCrew expert configures it to your firm
In your cloud (AWS, GCP or Azure) or in ours. Nothing shared with any other firm
Lowest-cost model that passes, with a spend limit per agent
Enterprise-grade security
The Agent Crew is your firm's set of PitCrew agents: 60+ ready-made agents for proposals, follow-ups, client requests, onboarding, fee checks, reconciliations, drift checks and reports, plus any agents you build. A PitCrew expert sets each one up to your rules, systems and approval path.
Yes. Describe the work in the PitCrew app or in Claude. PitCrew proposes the steps, the inputs and where a person signs off, then builds and tests the agent and adds it to your Agent Crew. Most firms start with us building the first agents, then their own teams build the next ones.
Branded pages you send to clients instead of a PDF. PitCrew fills in what it already knows from your CRM, the client completes the rest from one private link, and your team sees who needs you, who is filling it in and who is done. Firms use them for account opening and transfer forms, onboarding and KYC, and proposals.
You upload your billing manual, compliance manual and other procedures, and PitCrew turns each one into named rules with an owner and a version. Agents check their work against those rules, along with each client's signed documents and the regulations that apply to you, and tell you which rule decided what.
Yes. A person at your firm approves the key steps, and agents flag exceptions for review instead of acting on their own. Agents can update your CRM, custodian and billing systems, but only after sign-off, and every change is logged.
Every run is kept: what the agent read, which rules it checked, what it produced and who approved it. Those records are organized by category, time period and regulation, so an exam request can be answered from work already documented.
Yes. Ask Kitt answers questions on your runs, your documents and data, and the firm as a whole. Every answer cites the run, the page and the rule it came from.
Yes. PitCrew runs an MCP server that Claude connects to. Once connected, you can describe a new agent, have PitCrew build and test it, and start runs from the chat, with the result coming back to the conversation. You can also work with PitCrew from Slack, email and an API. Every run follows your firm's rules and is kept.
40+ systems, including custodians, CRMs, portfolio platforms, document stores, email and calendars, through standard APIs. Nothing gets replaced, and your firm's own APIs can be added so every agent can use them.
5 working days from data access to the first production run. A PitCrew expert configures PitCrew to your firm's rules, systems and approval paths.
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.