HIVEMIND DOCUMENTATION

Build AI Agents. Earn Passive Income.

WHAT IS HIVEMIND?

Hivemind is a platform for creating, sharing, and monetizing AI agents. Build custom AI assistants with unique personalities, publish them to the marketplace, and earn when others use your creations.

Build

Create agents with custom personalities and capabilities

Publish

Share your agents on the marketplace

Earn

Get rewarded when others use your agents

GETTING STARTED

1. Connect Your Wallet

Hivemind uses Solana wallet authentication. Connect your Phantom, Solflare, or any compatible wallet to get started.

Click "Connect" in the header → Select your wallet → Approve the connection

2. Explore the Marketplace

Browse existing agents created by the community. Add them to your workspace and start chatting immediately.

Browse Marketplace

3. Create Your First Agent

Use the Agent Builder to create custom AI agents with unique personalities, system prompts, and model configurations.

Open Builder

CREATING AGENTS

Agent Configuration

Name & Description

Give your agent a memorable name and clear description of what it does.

Personality Traits

Define traits like "helpful", "creative", "technical" that shape how your agent responds.

Tone & Verbosity

Choose between formal/casual tone and concise/detailed responses.

System Prompt

Write custom instructions that define your agent's behavior and expertise.

Model Settings

Available Models

Claude 3.5 Haiku - Fast and efficient
GPT-4o Mini - Balanced performance
Gemini 3 Flash - Quick responses

Temperature (0-2)

Lower = more focused, Higher = more creative

Max Tokens

Maximum length of agent responses (256-4096)

MARKETPLACE

Publishing Your Agent

1Complete your agent configuration in the Builder
2Select a category for your agent
3Click "Publish" and confirm
4Your agent is now live on the marketplace!

CHAT SYSTEM

Conversations

Chat with any agent in your workspace. Each conversation is saved and can be continued later.

• Add agents from the marketplace to your workspace

• Start new conversations with one or multiple agents

• View conversation history in the sidebar

• Delete conversations when no longer needed

SWARM MODE

Multi-Agent Collaboration

Swarm mode allows multiple agents to work together on a single task. Each agent contributes their unique perspective, and responses are synthesized into a comprehensive answer.

Example: Select a "Technical Writer" + "Code Expert" + "UX Designer" to get comprehensive documentation

To use Swarm mode, select 2 or more agents when starting a new conversation.

API REFERENCE

How Agent Requests Work

When you send a message to an agent, here's what happens behind the scenes:

1. Message Processing

POST /api/conversations/{id}/messages
Content-Type: application/json
x-wallet-address: {your_wallet}

{
  "content": "Your message to the agent"
}

2. System Prompt Construction

The agent's personality and system prompt are combined:

// System prompt is built from:
{
  base_prompt: agent.system_prompt,
  personality: {
    traits: ["helpful", "creative"],
    tone: "friendly",
    verbosity: "balanced"
  }
}

// Resulting prompt:
"You are an AI assistant with the following traits: 
helpful, creative. Respond in a friendly tone with 
balanced verbosity. {agent.system_prompt}"

3. AI MODEL REQUEST

Request is sent to the configured AI model:

// API call
POST /api/v1/chat/completions

{
  "model": "gpt-4o-mini",
  "messages": [
    { "role": "system", "content": "{constructed_prompt}" },
    { "role": "user", "content": "Your message" }
  ],
  "temperature": 0.7,
  "max_tokens": 1024
}

4. Response

// Response structure
{
  "userMessage": {
    "id": "msg_123",
    "role": "user",
    "content": "Your message",
    "created_at": "2026-01-20T..."
  },
  "agentMessage": {
    "id": "msg_124",
    "role": "agent",
    "content": "Agent's response...",
    "created_at": "2026-01-20T...",
    "agent": { "id": "...", "name": "Agent Name" }
  },
  "tokensUsed": 150
}

Swarm Mode Request

In Swarm mode, the task is sent to all agents in parallel:

POST /api/conversations/{id}/swarm
Content-Type: application/json
x-wallet-address: {your_wallet}

{
  "task": "Explain quantum computing"
}

// Response includes synthesized content from all agents
{
  "userMessage": { ... },
  "swarmMessage": {
    "id": "msg_125",
    "role": "system",
    "content": "Synthesized response from all agents..."
  },
  "synthesizedContent": "Combined insights from Agent1, Agent2...",
  "agentResponses": [
    { "agentId": "...", "response": "..." },
    { "agentId": "...", "response": "..." }
  ]
}