Hologram Developer
Documentation

Build, deploy, and run verifiable AI agents on the Hologram network. Concepts, tutorials, and reference for engineers.

Get started Read the introduction

Three paths

Learn the model. Build an agent. Run it anywhere.

Pick where you are on the journey. Each path links into a focused sidebar with everything in order.

Learn

The Hologram mental model — verifiable AI agents, the trust model, where each piece sits in the wider ecosystem.

Read the concepts

Build

Fork the starter, set an OpenAI key, chat with your own verifiable agent in 10 minutes. Then go deep on the Agent Pack — MCP, RBAC, approvals.

Start the quickstart

Run

Run locally with Docker Compose, or deploy to Kubernetes via the Helm chart used by every agent on vs.hologram.zone.

Deploy your agent

Recommended start

Fork, configure, ship.

Every Hologram agent is one YAML manifest — its Agent Pack — that captures personality, language, LLM, tools, authentication, and access control. Fork hologram-ai-agent-example, point at your LLM, and you have a running, trust-resolvable agent in ten minutes.

Open the quickstart

agent-pack.yaml

metadata:
  id: example-agent
  displayName: Example Agent

llm:
  provider: openai
  model: gpt-5.4-mini

mcp:
  servers:
    - name: context7
      url: https://mcp.context7.com/mcp
      accessMode: admin-controlled

flows:
  authentication:
    enabled: true
    credentialDefinitionId: ${CREDENTIAL_DEFINITION_ID}