LLMForce API (Beta)

This documentation describes the core API endpoints for the LLMForce platform.

Core API Endpoints

1. Search Expert

POST /v1/experts/search

Request

{
  "skills": ["string"],              // Required skills
  "task": {
    "description": "string",         // Task description
    "expected_duration": "string",   // e.g. "2h", "3d"
    "priority": "low" | "medium" | "high",
    "deadline": "string",           // ISO timestamp
    "type": "code_review" | "debugging" | "design" | "implementation" | "consulting",
    "artifacts": {                  // Any relevant links or files
      "repository": "string",
      "documentation": "string",
      "additional_context": "string"
    }
  },
  "requirements": {
    "min_experience_years": "number",
    "preferred_timezone": "string",  // Optional
    "language": ["string"],         // Required languages
    "max_budget": "number"          // Optional
  }
}

Response

{
  "experts": [{
    "expert_id": "string",
    "match_score": "number",
    "available_from": "string",     // ISO timestamp
    "estimated_completion": "string", // ISO timestamp
    "cost_estimate": {
      "amount": "number",
      "currency": "string"
    },
    "performance_metrics": {
      "completion_rate": "number",      // 0-1
      "on_time_rate": "number",        // 0-1
      "satisfaction_score": "number",   // 0-1
      "similar_tasks_completed": "number"
    },
    "availability": {
      "status": "available" | "busy" | "offline",
      "next_slot": "string",           // ISO timestamp
      "timezone": "string"
    }
  }]
}

2. Assign Task

POST /v1/experts/{expert_id}/assign

Request

{
  "task_id": "string",
  "task": {
    // Same as search task structure
  },
  "communication_preferences": {
    "update_frequency": "hourly" | "daily" | "on_completion" | "on_milestone",
    "notification_webhook": "string",  // URL for updates
    "preferred_format": "text" | "structured" | "both"
  }
}

Response

{
  "assignment_id": "string",
  "status": "accepted" | "rejected" | "pending",
  "expert_feedback": {
    "estimated_start": "string",    // ISO timestamp
    "estimated_completion": "string", // ISO timestamp
    "clarifying_questions": ["string"]  // Any immediate questions from expert
  },
  "communication": {
    "channel_id": "string",         // Unique communication channel ID
    "webhook_confirmed": "boolean", // Confirmation of webhook setup
    "update_schedule": {
      "first_update": "string",     // ISO timestamp
      "frequency": "string"         // e.g. "1h", "1d"
    }
  }
}

Expert Updates (Webhook Payload)

This is what the expert sends to the agent's webhook:

{
  "update_id": "string",
  "assignment_id": "string",
  "timestamp": "string",           // ISO timestamp
  "type": "progress" | "completion" | "blocker" | "question",
  "status": {
    "progress_percentage": "number",
    "current_phase": "string",
    "time_spent": "string",        // e.g. "1h 30m"
    "remaining_estimate": "string"  // e.g. "2h"
  },
  "content": {
    "message": "string",           // Update message
    "artifacts": {                 // Any new artifacts or updates
      "commits": ["string"],
      "documents": ["string"],
      "review_links": ["string"]
    },
    "questions": [{               // Any questions requiring agent's input
      "question_id": "string",
      "question": "string",
      "blocking": "boolean",      // Whether this blocks progress
      "options": ["string"]       // Possible answers if applicable
    }]
  },
  "next_update": "string"         // ISO timestamp for next scheduled update
}

Agent Response to Expert Questions

POST /v1/assignments/{assignment_id}/respond

{
  "question_id": "string",
  "response": "string",
  "additional_context": "string",
  "artifacts": {                // Any additional resources
    "links": ["string"],
    "content": "string"
  }
}