āļāļģāļāļāļīāļāļēāļĒ
AI Trading and Advisory Assistant, is a cTrader cBot that uses machine learning and current large language models (LLMs) to analyze live market data, provide trading advice, or execute trades automatically.
The latest release connects through OpenRouter, NVIDIA, and Router9. These providers give you access to many model families through a single, OpenAI-compatible API, including models from OpenAI/ChatGPT, Anthropic Claude, Google Gemini, DeepSeek, Meta, Mistral, Qwen, Kimi, and other providers as they become available. This includes new-generation models such as Anthropic Claude Fable 5, ChatGPT 5.6 Sol, Qwen 3.8, and Kimi K3 when the selected provider lists them and exposes a chat-completions-compatible model ID.
Model availability, names, pricing, context limits, and free-tier access can change. Always copy the exact model ID from the selected provider's current model catalog.
How TraderAI Works
TraderAI sends selected market data to your chosen AI model in real time. The model follows your custom prompt and returns either advice or structured trading actions. This removes the need to copy market data manually or execute every decision yourself.
AI can use price action, candlesticks, tick data, market depth, chart indicators, open positions, and recent decision history to identify trends and evaluate possible trades. AI output can be inaccurate; test every strategy on a demo account before using it with real funds.
Two Operating Modes
Advisor mode analyzes the supplied market state and displays advice. You remain responsible for deciding whether to trade.
Trader mode allows the model to open, close, or modify positions on the selected trading account. Built-in execution limits validate model actions before they reach the account.
Supported AI Providers
OpenRouter
OpenRouter provides access to many commercial and open models through one API key and model naming scheme.
Model catalog: openrouter.ai/models
Default model ID: openai/gpt-oss-20b:free
Enter the exact catalog ID, usually in publisher/model-name format.
NVIDIA
NVIDIA provides hosted NVIDIA NIM endpoints for supported text-generation and reasoning models.
Model catalog: build.nvidia.com/models
Select a model with a Free Endpoint or another hosted API endpoint.
Open the model page and copy the exact API model ID. Not every catalog entry is a chat model; image, audio, embedding, and downloadable-only models are not suitable for this cBot.
Router9
Router9 routes supported models from multiple upstream providers through one endpoint.
Model catalog: router9.com/models
Enter the exact value shown in the catalog's Model ID column.
The catalog identifies the current models, upstream providers, context limits, prices, and free-tier availability.
cBot Parameters
AI Config
AI Role â Select Advisor or Trader.
AI Provider â Select OpenRouter, Nvidia, or Router9.
AI Temperature â Controls response variation. The default is 0.2.
- 0.1â0.3: more consistent and conservative
- 0.4â0.7: balanced
- 0.8â1.0: more varied and creative
Lower values are generally better for repeatable trading decisions. Some providers or reasoning models may ignore or restrict temperature.
Execute On - Select when the cBot requests a new decision:
TrendBar: when a new chart-timeframe bar is available
TickBar: from incoming tick activity, subject to the decision interval
Decision Interval (Seconds) - Minimum time between AI requests. The default is 30. A larger value reduces API usage and helps avoid provider rate limits.
API Timeout (Seconds) - Maximum time to wait for a provider response. The default is 45 seconds.
History Decisions - Number of recent decisions supplied to the model, from 0 to 20. The default is 5.
Prompt - Defines the strategy, constraints, and behavior expected from the model. Create and validate prompts in a demo environment. The prompt does not override the cBot's execution firewall.
Provider Config
Provider API Key - Enter the API key issued by the provider selected under AI Provider.
Provider Model Name - Enter the provider's exact model ID. Model IDs are case-sensitive and provider-specific. Do not enter a ChatGPT product or subscription name unless it appears as an API model ID in the selected provider's catalog.
Example OpenRouter setup
AI Provider: OpenRouter
Provider Model Name: moonshotai/kimi-k3 or anthropic/claude-fable-5 or nvidia/nemotron-3-ultra-550b-a55b:free
Provider API Key: your OpenRouter API key
When switching providers, replace both the API key and model ID. A direct OpenAI, Gemini, Anthropic, or DeepSeek key from an older TraderAI release is not valid unless the selected routing provider explicitly supports using that key.
Model Input:
Include Candlesticks - Include chart candlestick data in the model input.
Candlesticks Window - Number of recent candlesticks to include.
Include TickBars - Include recent bid/ask tick data.
TickBars Window (Minute) - Number of minutes of tick history to include.
Include MarketDepth - Include the current market-depth snapshot. Market depth may not be available during backtesting.
Include Chart Indicators - Include supported indicators attached to the chart.
Indicator Values Window - Number of recent values to include for each chart indicator, from 1 to 100.
Larger input windows consume more tokens and can increase latency and cost. Some combinations of timeframe and backtesting data source may not provide tick or depth data.
Trade Size
Min Trade Size (Lots) - Smallest trade size the AI is allowed to request.
Max Trade Size (Lots) - Largest trade size the AI is allowed to request.
Execution Firewall
These limits apply in Trader mode and help constrain AI-generated actions:
Max Exposure (Lots) - Maximum total exposure the cBot may create.
Max Bot Positions - Maximum number of positions managed by the cBot.
Max Actions Per Decision - Maximum number of open, close, or modify actions accepted from one AI response.
Max Spread (Pips) - Prevents new entries when the current spread exceeds the configured value.
Max Drawdown (%) - Prevents additional risk when the configured drawdown limit is reached.
These controls reduce risk but cannot eliminate losses, provider errors, latency, slippage, or incorrect model output.
Bring Your Own Key (BYOK)
TraderAI does not bundle provider access. Create an account with your preferred provider, obtain an API key, and fund or configure that account as required:
- OpenRouter: https://openrouter.ai/
- NVIDIA Build: https://build.nvidia.com/
- Router9: https://www.router9.com/
Never share an API key or place it in screenshots, public source code, or support messages. Provider usage charges and limits are controlled by the provider, not by TraderAI.
Choosing a Model
For trading or advisory use, choose a model that supports text chat completions, follows structured instructions reliably, and has enough context for your selected market-data windows. Compare latency and cost on a demo account. A larger or reasoning-focused model is not automatically a better trading model.
Examples of model families that may be available through the supported providers include:
- Anthropic: Claude Fable 5
- OpenAI/ChatGPT: ChatGPT 5.6 Sol
- Alibaba Qwen: Qwen 3.8
- Moonshot AI: Kimi K3
These are display names, not guaranteed API identifiers. Copy the exact model ID from the provider catalogâfor example, a provider may add an organization prefix, use lowercase characters, or publish a versioned suffix. If a model is absent from the selected provider's catalog, it cannot be used through that provider yet.
Use these live catalogs for the current list:
NVIDIA: build.nvidia.com/models
OpenRouter: openrouter.ai/models
Router9: www.router9.com/models
Important Notice
TraderAI is a technical tool, not financial advice or a guarantee of performance. AI-generated analysis and trades can be wrong and can cause substantial losses. Test with a demo account, use conservative limits, monitor the cBot, and ensure automated trading is permitted in your jurisdiction and by your broker.
āļŠāļĢāļļāļ
The product supports two operational modes:
1. **Trader Mode**: Executes trades automatically on the userâs selected trading account based on custom prompts.
2. **Advisor Mode**: Provides trading advice based on market analysis, leaving trade execution to the user.
Users can customize prompts to tailor AI behavior according to their trading strategies. The system supports multiple AI models, including the latest OpenAI models (e.g., GPT-4 variants), all Gemini models, DeepSeek chat and reasoner models, and Claude models, with Grok and Mistral models planned for future integration.
Key configurable parameters include AI role, model selection, reasoning activation, AI temperature (controlling creativity vs. determinism), maximum token usage, execution timing (market trend bars or tick bars), and market data inputs such as candlesticks, tickers, and market depth. Trade size limits can also be set.
Access to AI models requires user-provided API keys. Sample prompts demonstrate strategies for high-frequency trading, momentum-based trading, and statistical arbitrage across forex, commodities, crypto, and equity markets.
āļĢāļĩāļ§āļīāļ§āļāļēāļāļĨāļđāļāļāđāļē
5 | 33 % | |
4 | 67 % | |
3 | 0 % | |
2 | 0 % | |
1 | 0 % |
