Skip to main content

Advanced Features

This guide dives deep into OpenHuman's advanced features to help you fully leverage your agent's capabilities.

Multi-modal Interaction​

Image Processing​

const response = await client.agents.run(agent.id, {
  message: 'Describe this image',
  attachments: [
    {
      type: 'image',
      url: 'https://example.com/image.jpg',
    },
  ],
});

File Processing​

const response = await client.agents.run(agent.id, {
  message: 'Summarize this PDF',
  attachments: [
    {
      type: 'file',
      name: 'report.pdf',
      url: 'https://example.com/report.pdf',
    },
  ],
});

Local AI​

Run AI models on your machine, keeping data private:

[local_ai]
enabled = true
provider = "ollama"  # or "lm_studio"
ollama_url = "http://localhost:11434"
embedding_model = "all-minilm:latest"
memory_model = "gemma3:1b-it-qat"

Three presets available:

  • Embeddings Only - Local vector computation
  • Memory + Reflection - Local processing of memory and reflection
  • Fully Local - All workloads run locally

Custom Tools​

Register an HTTP Request Tool​

const httpTool = await client.tools.register({
  name: 'http-request',
  description: 'Send an HTTP request',
  parameters: {
    type: 'object',
    properties: {
      url: { type: 'string' },
      method: { type: 'string', enum: ['GET', 'POST'] },
    },
  },
  execute: async ({ url, method }) => {
    const response = await fetch(url, { method });
    return await response.json();
  },
});

Register a Database Tool​

const dbTool = await client.tools.register({
  name: 'query-db',
  description: 'Query a database',
  parameters: {
    type: 'object',
    properties: {
      sql: { type: 'string' },
    },
  },
  execute: async ({ sql }) => {
    const result = await db.query(sql);
    return result;
  },
});

Agent Chain Invocation​

const chain = await client.chains.create({
  name: 'Research Writing Chain',
  steps: [
    {
      type: 'agent',
      agentId: researchAgent.id,
      prompt: 'Research topics related to {{input}}',
    },
    {
      type: 'agent',
      agentId: writerAgent.id,
      prompt: 'Write an article based on the following research: {{previous}}',
    },
  ],
});

const result = await client.chains.run(chain.id, {
  input: 'Trends in AI Agent Development',
});

Remote Core​

Deploy the Rust core on a remote server and connect via RPC:

[core]
mode = "remote"
rpc_url = "https://your-core.example/rpc"
token = "your-core-token"

Ideal for teams sharing the same core instance, or accessing from different network environments.

Webhook Callbacks​

await client.agents.update(agent.id, {
  webhook: {
    url: 'https://your-server.com/webhook',
    events: ['onComplete', 'onError'],
  },
});

Next Steps​