botreeveJoin the waitlist

How it works

Six steps. A model helps you write your policy, ahead of time. Once a bot proposes something, the rest is either you, or code that cannot be talked into anything.

  1. 01
    Connect a bot you already use

    Sign in to botreeve, then connect a bot or agent from a service you already trust — your own, or one shared with you inside a group. The first connector is trading, through a dedicated Robinhood account on Robinhood's own Agentic Trading program. botreeve stores an encrypted, per-tenant credential for that connection and nothing more.

  2. 02
    You write the rules, in your own words

    Every rule a bot acts on is one you wrote yourself, in plain language, or an alert from a system you already subscribe to independently of botreeve. The Chief of Bots compiles what you wrote into a policy — a machine-checkable set of limits and permissions — before any bot has anything to propose, and you can always read back exactly what your words compiled to. botreeve does not author rules, rank instruments, or suggest what to do next.

  3. 03
    The bot proposes, continuously

    Once connected, a bot evaluates your rules around the clock and proposes an action when one fires — even while your phone is off. Proposing costs nothing and does nothing.

  4. 04
    The Chief of Bots tells you what's waiting

    botreeve's conductor agent knows every bot's state across every connection — yours and, inside a group, the ones your friends or family run. It surfaces the one thing that actually needs you instead of a feed you have to scroll, and answers plain-language questions about your fleet.

  5. 05
    When your policy calls for your tap

    Some rules end in “ask me first.” When one does, the proposal reaches your Lock Screen with two options, deliberately unequal. Declining is the safe path, so “Not now” resolves right there in one tap. “Authorize” opens the app to the full proposal instead of executing anything — a Lock Screen band can't show your limits or what's left of today's cap, and authorizing something without that context is exactly the failure this design avoids.

  6. 06
    A deterministic gate enforces your policy

    Whether a proposal held for your tap or acted the moment it fired — your call, made when you wrote the rule — the same separate piece of code checks it against the policy compiled from your rules and carries it out. That code is pure and synchronous: same input, same result, every time. The full breakdown is below.

The gate: you write the rules, the Chief compiles them, code enforces them

A bot that can talk its way into an action is a bot that can talk its way past a limit. botreeve keeps the part that talks and the part that acts as two different pieces of code — and by the time an action is actually on the table, no model is in the room.

your words   →  compile()  →  policy    the only place a model touches this
a proposal  →  decide()    →  act()      no model in the path, ever

You write your rules in plain language. The Chief of Bots compiles those words into a policy — ahead of time, before any bot has anything to propose — and you can always read back exactly what your words compiled to. From there, whether a proposed action is inside that policy is decided by a separate function that is pure and synchronous: no network calls, no memory of the conversation, the same answer for the same input every time. The model that helped you write the policy is not in the room when a bot proposes an action, and it never gets a vote on whether that action goes through.

What your policy says is entirely up to you. One rule might hold every proposal for your tap. Another might say act the moment it fires, up to a limit you set, because you trust the rule you wrote for it. botreeve does not decide which — you do, in the words you write, and you carry the consequences of what you authorize.

That gate applies to any consequential action a connected bot might take — spending money, posting something publicly, sending a message on your behalf, booking a reservation. Trading is one instance of it, not the reason it exists.

Nothing acts outside the policy you authorized, and you can always read exactly what you authorized — the specific action, the rule that produced it, and whether that rule held it for your tap or let it run — inside the app.

botreeve carries no trading signals and makes no recommendations, at any step above. Read the full risk disclosure before connecting a trading account.