Pluto AI Agent
Use natural language to work through strategy, backtesting, trading bot, and market-related tasks. AI Agent understands intent, offers guidance, and operates within clearly defined safety boundaries.
With private deployment, your data and conversation context stay in your own environment. The underlying trading engine still enforces all risk and permission checks, so AI cannot bypass limits or safeguards.
AI + Trading Dual Engine
Pluto uses a dual-layer architecture of "AI Agent + Trading System Layer": AI Agent serves as the "brain," understanding natural language, providing analysis and suggestions, and orchestrating capabilities across strategy, backtesting, live bots, market data, analysis, and system configuration. The underlying trading system handles execution and strictly enforces risk control and permission checks—AI cannot bypass position limits, stop-loss protection, or unauthorized fund operations.
AI Agent Decision brain
Interprets natural language, provides analysis and suggestions, and intelligently orchestrates system tools.
Trading engine layer Secure execution
Handles underlying risk validation and real order execution, ensuring fund and permission safety.
AI Agent (Brain)
Intent understanding / Strategy analysis / Task orchestration
Trading engine layer (Execution)
Managed Through Conversation
Describe what you want in natural language. AI Agent can understand the request, offer suggestions, and help execute tasks within your permission and risk boundaries. Strategy setup, backtesting, bots, market data, and system configuration can all be handled through conversation.
Strategy
Help with strategy design, entry and exit logic, explanations, and trade-offs; create new strategies and take you directly to the configuration page.
Backtest & optimization
Run backtests, explain metrics such as return, drawdown, and Sharpe ratio, and suggest improvements to parameters and configuration.
Live bots
Create and configure bots, explain status, query orders and positions, and troubleshoot why positions are not opening.
Market & positions
Query market prices and position data, interpret trends with indicators, and flag potential risk points.
System usage
Guide you through exchange setup, notification configuration, and first-time workflows such as your first strategy or first backtest.
Multi-Model Support
Choose a provider and enter your API key and model name in AI config. You can select based on cost, speed, quality, or network conditions, and set primary and backup models for automatic failover.
You can also configure temperature, max tokens, timeout, and streaming. If you use a self-hosted endpoint or proxy, enter the Base URL. See the in-app settings for full details.
OpenAI
GPT series
Claude
Anthropic
DeepSeek
Cost-friendly
Groq
Low-latency inference
Gemini
Grok
xAI
From conversation to execution
From one sentence to in-system execution: understand → orchestrate → execute (or confirm) → iterate.
Describe your goal in natural language
e.g. "Create a BTC MACD trend strategy" or "Run 30-day backtest on 1H."
AI Agent parses and orchestrates
AI Agent interprets your intent, selects the right tools for strategy, backtesting, or data queries, and presents intermediate results in tables or cards when helpful.
Engine executes by rules
Orders and configuration changes are carried out by the underlying engine. Any invalid or out-of-bounds action is blocked, and important steps require your confirmation.
Review results and follow up
Refine parameters based on backtest reports and runtime logs. You can also continue the workflow through Telegram if notifications and smart replies are configured.
Disclaimer：AI Agent provides analysis, suggestions, and operational support only—no profit guarantees. Not investment advice. Assess risk before going live.
Your own AI Agent
Use AI to access knowledge, market data, and much of the system's operational workflow through conversation, with risk controls and permission safeguards built in.