Skip to main content
HyperAI is Hyperoru’s conversational AI assistant. It helps you create strategies, analyze markets, manage traders, and interact with the platform through natural language.

​
Overview

HyperAI is more than a chatbot — it’s an agent system with persistent memory, specialized skills, tool calling, and the ability to orchestrate sub-agents for complex tasks.

​
LLM provider configuration

HyperAI uses the system-wide LLM settings by default but can be overridden per conversation: 
curl -X POST https://api.hyperoru.com/api/hyper-ai/chat \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Analyze the current BTC market conditions",
    "model": "claude-sonnet-4-20250514",
    "api_key": "sk-ant-your-key",
    "base_url": "https://api.anthropic.com/v1"
  }'
If no model override is provided, HyperAI uses HYPER_INSIGHT_SERVICE_TOKEN and HYPER_INSIGHT_API_BASE_URL from environment variables.

​
Conversation management

Conversations are persisted and can be resumed across sessions: 
curl -X POST https://api.hyperoru.com/api/hyper-ai/conversations \
  -H "Content-Type: application/json" \
  -d '{"user_id": "your-user-id", "title": "BTC Strategy Research"}'

​
Memory system

HyperAI maintains long-term memory organized into four categories:
CategoryPurposeExample
PreferencesUser trading preferences and risk tolerance”Prefers 3x max leverage, only trades BTC and ETH”
DecisionsPast trading decisions and their outcomes”Shorted BTC at 65k based on RSI divergence — resulted in 3% profit”
LessonsLearned patterns and mistakes”Avoid entering positions during low-volume Asian sessions”
InsightsMarket observations and analysis”BTC tends to follow ETH breakouts with a 2-4 hour delay”
Memories are automatically extracted from conversations and recalled when relevant to future queries.

​
Skill system

HyperAI has 9 built-in skills that handle specialized tasks:
SkillCapability
Market AnalysisAnalyze price action, indicators, and market regime
Strategy CreationGenerate prompt templates or program code from descriptions
Signal DesignCreate signal definitions and pools from natural language
Backtest RunnerConfigure and execute backtests, interpret results
Trade ReviewAnalyze recent trades and suggest improvements
Risk AssessmentEvaluate portfolio exposure and position sizing
News BriefingSummarize relevant news and sentiment
Factor AnalysisEvaluate quantitative factors and their effectiveness
Platform HelpGuide users through platform features and API usage
Skills are automatically selected based on the user’s intent.

​
Tool system

HyperAI uses function calling to interact with the platform programmatically. When you ask it to create a trader, run a backtest, or check positions, it calls the appropriate API tools internally. 
User: "Create a new trader called ETH Scalper with 10x max leverage"

HyperAI: I'll create that trader for you.
→ [Calls: create_trader(name="ETH Scalper")]
→ [Calls: update_trader_settings(max_leverage=10)]

Done! Created trader "ETH Scalper" (id: abc123) with 10x max leverage.

​
Sub-agent orchestration

For complex tasks, HyperAI spawns specialized sub-agents that work in parallel:
1

Task decomposition

HyperAI breaks a complex request into sub-tasks (e.g., “Build me a complete BTC trading setup” becomes: create trader, write strategy, define signals, create binding).
2

Agent delegation

Each sub-task is assigned to a specialized agent with the appropriate skill and context.
3

Result synthesis

Sub-agent results are collected and synthesized into a coherent response with a summary of all actions taken.
Sub-agent orchestration happens transparently. From the user’s perspective, HyperAI handles the full request and reports back with results.