MCP App Store
by KeenEthics
Tally Markets icon

Tally Markets

By Prosperity Labs, LLC

Get New App Alerts

Overview

Every trading connector makes your AI more capable. Tally makes it accountable. Your AI is the analyst; Tally is the risk officer. Tally gives ChatGPT a trading journal it can operate but never rewrite: every trade starts as a pre-registered thesis with exit rules written before entry, a 24-hour cooling-off separates conviction from capital, and rule-breaking exits require a written justification that stays on the record forever. Ask for a checkup and ChatGPT briefs you on every active thesis, logs fresh evidence against your exit criteria, and flags anything that has objectively triggered. Every closed trade mints a public receipt — entry, exit, result, and your reasoning, word for word. Tally never executes trades.

Tools

arm_thesis

ChatGPT
Arm a thesis after its cooling-off has elapsed — the gate between registration and entry eligibility. One-way: an armed thesis cannot return to cooling-off.

close_thesis

ChatGPT
Close a thesis via one of the exit paths. reason 'discretionary' REQUIRES override_reason (the user's written justification) — it is logged to the override ledger before the close. All positions must already be closed.

create_thesis

ChatGPT
Pre-register a thesis. Criteria are immutable afterward. Armed at once by default; pass cooling_off: true for a 24h cooling-off before arming. Requires >=2 invalidation and >=1 dissemination criteria — each an observable fact, not a price level.

declare_mandate

ChatGPT
Declares what the agent is supposed to do, as structure: asset classes, the largest order as a share of the account or in dollars, cadence, hours, usual symbols. Tally runs the mandate from the first scan beside the learned baseline and never loosens it. A change to an existing mandate raises a notice to the owner; declare it once and honestly.

fire_criterion

ChatGPT
Mark a thesis criterion as FIRED with the evidence that fired it. One-way: a fired criterion stays fired and cannot be unfired through this API.

get_checkup

ChatGPT
The morning briefing. Returns monitoring first — the status word per account (Learning, Normal, Elevated, High), orders in the last 24 hours, anomaly alerts of the last 7 days with whether the user answered them, the broker connection — then needs_attention (unanswered anomalies, a connection that stopped reading, fired criteria, due or approaching time exits, windows about to close, theses ready to arm), then every active thesis with its deadlines and criteria state.

get_execution_evidence

ChatGPT
Read archived order observations, exact intent attribution, individual fills and agent decisions for a Tally connection. Values are recorded observations, not a live protection check. Does not place orders or change trading state.

get_started

ChatGPT
Read this first: Tally's discipline protocol and how the tools fit together.

get_thesis

ChatGPT
One thesis in full: statement, mechanism, criteria (with fired state), positions, review.

list_theses

ChatGPT
All of the user's theses, newest first, any status.

log_agent_activity

ChatGPT
Records what the agent researched, considered, or proposed, as symbols and kinds only; no text is accepted or stored. Tally's monitoring reads the reports beside the account's actual orders, and the status screen counts them and names any symbol ordered without a report.

record_reading

ChatGPT
Log a data reading against a thesis — the generic sensor primitive. Records whatever value a criterion names (a CoT index, a premium, a calendar fact) with its source and timestamp, building the thesis's durable evidence trail.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps