Overview

The Target Generator Service is a specialized service designed to generate structured AI-achievable targets based on mission or milestone descriptions using LLM technology. The service operates in two modes (target and milestone) to support hierarchical planning, ensuring all generated targets are verifiable through message or state analysis.

• Mission-Based Generation: Generate targets based on mission descriptions
• Dual Mode Operation: Support for both target and milestone generation
• Dependency Management: Automatic creation of logical target dependencies
• Verifiable Outcomes: All targets include clear completion conditions
• Context Awareness: Considers existing targets when generating new ones
• AI-Centric Design: All targets designed to be achievable through AI capabilities
• Automatic Validation: Ensures outputs follow required schema and naming conventions

Traditional task breakdown often involves manual task decomposition, static dependency mapping, and fixed completion criteria. The Target Generator Service improves this by using LLMs for intelligent target generation, creating dynamic dependency chains, generating verifiable completion conditions, and maintaining context awareness.

Endpoints

Process Endpoint

The process endpoint performs target and milestone generation based on mission descriptions.

API Endpoint

Endpoint Parameters

Request

content (string, required)

The mission or milestone description to generate targets from.

mode (string, required)

Target generation mode:

• target: Generate specific, actionable targets from a milestone
• milestone: Generate strategic milestones from a mission

context (object, optional)

Optional existing targets and additional context:

• existingTargets (object): Contains todo and done targets
  • todo (array): List of incomplete target objects
  • done (array): List of completed target objects

config (object, optional)

Model and generation configuration:

• modelId (string): Model to use for generation (default: "mistral-large-2411")
• temperature (number): Controls generation creativity (default: 0.1, range: 0.0-1.0)
• maxTokens (integer): Maximum tokens for model generation (default: 2000)
• targetCount (integer): Number of targets/milestones to generate (default: 5, range: 1-20)

Response

status (string)

Status of the processing ("success" or "error").

result (object)

For successful responses:

• todo (array): List of generated targets
• done (array): List of completed targets (if provided in context) Each target includes:
  • id (integer): Target identifier
  • target (string): Target description
  • done_when (string): Completion criteria
  • depends_on (array): Dependencies

metadata (object)

Processing metadata:

• model (string): Model used
• created_at (string): Creation timestamp
• processed_at (string): Processing timestamp
• instance_id (string): Instance identifier
• tracking_id (string): Unique tracking ID
• mode (string): Generation mode used ("milestone_generation" or "target_generation")

Examples

Request

Mission to Milestones (Milestone Generation Mode)

Milestone to Targets (Target Generation Mode)

Complete Project Breakdown Example - Step 1

Complete Project Breakdown Example - Step 2

Target Generation with Custom Configuration

Milestone Generation with Dependencies

Response

Milestone Generation Response

Target Generation Response

Error Response

Generation Modes

The service supports two distinct modes for hierarchical planning:

Target Mode

• Generates specific, actionable targets
• Focuses on concrete tasks
• Includes detailed completion criteria
• Suitable for immediate execution

Milestone Mode

• Generates high-level strategic targets
• Focuses on major project phases
• Includes broader completion criteria
• Suitable for project planning

Hierarchical Planning Approach

The service implements a hierarchical planning approach:

Mission Level

• Top-level objective or goal
• Example: "Build a scalable e-commerce platform"
• Used as input for milestone generation

Milestone Level (Generated in Milestone Mode)

• Strategic phases or major achievements
• Derived directly from the mission
• Broader in scope than targets

Target Level (Generated in Target Mode)

• Specific, actionable tasks
• Derived from individual milestones
• Concrete completion criteria

Best Practices

Mission Description

• Be specific about desired outcomes
• Include key requirements
• Specify constraints
• Provide relevant context

Mode Selection

• Use target mode for specific tasks
• Use milestone mode for project planning
• Match mode to abstraction level
• Consider dependency complexity

Configuration Tuning

• Adjust target count based on scope
• Use low temperature for consistency
• Set appropriate token limits
• Include relevant context

Iterative Planning

• Start with milestone generation
• Generate targets for highest priority milestone
• Update and adjust based on progress
• Maintain dependency consistency

Limitations

• Fixed generation modes
• Text-based generation only
• Fixed dependency types