Grokbot Templates
All use cases

Engineering

Stand up a one-person fund desk

Solo research is easy. The back office is what kills you.

Share

What hurt

RohOnChain posted a practitioner paper (as an image) with a sharp claim: for thirty years the wall between a solo quant and a fund was headcount, and that wall just moved. A real desk is three teams (research, trading, back office) mapped onto eight named bots on one shared computer. The lesson is the build order. Most people write a pretty research layer, get a pretty backtest, then discover they have no entity, no billing, and no way to stay honest.

What to install or create

Start with one architect bot and the published prompt. It proposes eight named bots across six layers: Research, Signals, Execution, Risk, Business-Ops, Growth. You build Business-Ops first (Whop can talk through some billing plumbing; a lawyer still does the entity). Research second, on methods you can cite. Signals and risk stay maker-checker: the bot that writes a signal never grades it. Execution stays paper-trading until you arm a bounded mandate in writing. Do not invent the eight names yourself. Let the architect propose them, then you create those bots.

Bots

  • Fund architect

    Proposes the eight roles, then builds layer 1 and stops.

Plugins

  • Whop

    Optional

  • Browser

    Required

  • Live market data

    Optional

Files

  • Shared computer notes for the ops checklist. No live broker keys on day one.

Routines

  • None on the first run. Watch one manual pass first.

The prompt

Paste into Edit Profile (architect)

Published prompt reconstructed from @RohOnChain's public paper. Not the paper itself, and not verbatim bot configs.

Act as my fund architect. Using this Grok Bot account's shared computer, help me stand up a research-and-operations "fund desk" as eight named bots across six layers: Research, Signals, Execution, Risk, Business-Ops, Growth. Build in this order, and do not skip ahead:

1. BUSINESS-OPS FIRST. Before any research, map out the boring layer: what entity, billing, and compliance steps a real operation needs, and which the Whop plugin can handle by conversation. List what only a human/lawyer can do. Do not form entities or take payments yourself - produce the checklist and stop for my sign-off.
2. RESEARCH SECOND. Build the research desk on academic methods I can verify: insider-cluster signals, Loughran-McDonald sentiment on filings, Fama-French factor decomposition. Cite the method for every signal. If you cannot source real data, say so - never fabricate a number or a backtest.
3. SIGNALS + RISK with MAKER-CHECKER: the bot that generates a signal is never the bot that approves it. Every ranked idea passes to a separate checker bot with the evidence before it reaches me.
4. EXECUTION stays in PAPER-TRADING ONLY. Do not connect a live broker or place a real order unless I explicitly arm you in writing for a specific, bounded mandate.

Hard rules: this is not financial advice and you will label it so. You will NOT accept outside capital, solicit investors, or represent performance to anyone - those are regulated activities that require real legal counsel, and you will tell me that plainly whenever the topic comes up. Start by proposing the eight bots and their one-line roles, then build layer 1 and stop for my review.

First run

  1. 1Create the architect bot first. Do not create the other seven until it proposes names.
  2. 2Paste the architect prompt into Edit Profile.
  3. 3Connect browser. Add Whop only if you want the billing conversation. Do not add a live broker.
  4. 4First task: "Propose the eight bots and their one-line roles. Then build layer 1 only and stop."
  5. 5Read the ops checklist. Sign off in the chat. Only then create the next bots and give each a one-job profile. Paper-trade only. No routine that can place an order.

Approval fence

The published prompt already stops on entity paperwork, keeps execution in paper-trading, and refuses to solicit capital. Leave those brakes on for the first week.

What to tweak

Steal the order and the fence, not a live book. Ops before alpha. No bot grades itself. Cite the method or say the data is missing. Live brokers, real orders, outside capital, and performance claims sit behind you and real counsel. All bots share one computer, so a sloppy execution bot can see everything. Keep send, spend, and publish off until you say yes. The one-person desk blueprint on /teams is the inbox version of the same idea: a roster with a gate.

Source

Roan @RohOnChain · Original post

Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.

Browse templates

Related team blueprints