Task

Thoughtful Task Loop

Advanced LLM orchestration with meta-cognition capabilities.
Task provides intelligent model rotation, self-reflection, and dynamic weighted selection based on performance across multiple AI providers.

Simple Task orchestration

import { mindTask } from "@just-every/task";

// Run an agent with meta-cognition
const result = await mindTask({
  agent: { name: 'MyAgent' },
  task: 'Analyze this code for performance issues',
  runAgent: async (agent, input, history) => {
    // Your LLM call logic here
    const response = await callYourLLM(input, history);
    return { response };
  }
});

// Task handles intelligent model rotation and self-reflection

Advanced orchestration features

// Task provides intelligent features out of the box:

// 1. Ensemble Model Management
//    Multiple AI models work in parallel or sequence

// 2. Chain-of-Thought Reasoning
//    Maintains connected reasoning thread across interactions

// 3. Meta-Cognition
//    System periodically analyzes and adjusts its own performance

// 4. Dynamic Model Hierarchy
//    Weighted model selection based on historical performance

// 5. Automatic cost tracking and conversation history

Core Capabilities

Intelligent Model Rotation

Automatic switching between AI providers based on performance.

Periodic Self-Reflection

System analyzes and adjusts its own performance over time.

Thought Management

Chain-of-thought reasoning with intelligent pacing.

Multi-Provider Support

Works with OpenAI, Anthropic, and other LLM providers.

npm i @just-every/taskRead the spec

← Ensemble Magi →