Optional AI evaluation

Add a five-agent review when a signal needs another check.

Pro and Advanced plans can place an AI committee between parsing and execution on selected channels.

Optional AI evaluation illustration

See it in the product.

Technical Analyst

Momentum and ATR aligned

PASS

Channel Analyst

History consistent

PASS

Signal Supporter

Case for entry built

FOR

Signal Critic

Spread risk at session open

FLAG

Signal Evaluator verdict

Confidence 74

EXECUTE
Technical and channel analysis

One agent reads live market conditions while another weighs the channel's tracked history and recent accuracy.

Supporter and critic

A supporter builds the strongest case for the trade and a critic argues the risks against it, so both sides are heard.

Final evaluator decision

A fifth agent weighs the debate and returns EXECUTE, REDUCE_SIZE, or REJECT with a confidence score.

Optional and per-channel

Turn the committee on only for channels that need a second opinion, and leave fast, direct copying on the rest.

How it works.

1

Enable it on a channel

On Pro or Advanced, switch AI evaluation on for the channels where you want a second opinion.

2

Agents review the parsed signal

After parsing, the technical, channel, supporter, and critic agents each assess the trade.

3

Evaluator returns a decision

The evaluator combines the views into EXECUTE, REDUCE_SIZE, or REJECT, within your daily evaluation allowance.

4

Execution follows the verdict

Approved trades continue to sizing and the guard; rejected trades are skipped and logged for review.

Research foundation

TradingAgents: Multi-Agents LLM Financial Trading Framework

Yijia Xiao, Edward Sun, Di Luo, Wei Wang · arXiv:2412.20138

The committee design is inspired by the TradingAgents framework, which structures trading decisions as a debate between specialized LLM agents: analysts assess conditions, opposing researchers argue the bull and bear case, and a final layer weighs both sides before acting. Our five-agent review applies the same adversarial structure to a single Telegram signal: two analysts, a supporter, a critic, and an evaluator that settles the debate.

Read the paper on arXiv

Frequently asked.

Is AI evaluation mandatory?

No. AI evaluation is optional and off by default. Starter uses direct auto-execution, and Pro and Advanced can enable the committee per channel.

What are the five agents?

Technical Analyst, Channel Analyst, Signal Supporter, Signal Critic, and Signal Evaluator. The first four assess the trade and the evaluator makes the final call.

What decisions can it return?

The evaluator returns EXECUTE, REDUCE_SIZE, or REJECT with a confidence score. Pro includes up to 30 evaluations per day and Advanced up to 80.

Build the workflow around your rules.

Start the 3-day trial for $1. Configure your channels, risk controls, and execution preferences from the guided setup.

Start for $1