{
  "schema_version": "agentpmt.docs.v1",
  "generated_at": "2026-07-26T01:18:34.871Z",
  "release_id": "9bd28307d3dc8ff05bcac2f755a96681a9e650091caba4ab5ecbd5a1d6d62acc",
  "id": "69f31f72ad2e1817f7ace932",
  "page_url": "https://www.agentpmt.com/docs/core-concepts/agents",
  "markdown_url": "https://www.agentpmt.com/docs/core-concepts/agents?format=agent-md",
  "json_url": "https://www.agentpmt.com/docs/core-concepts/agents?format=agent-json",
  "source_path": "content/docs/agents.mdoc",
  "updated_at": "2026-07-23T01:20:26.647Z",
  "headings": [
    {
      "depth": 2,
      "id": "how-agents-connect",
      "title": "How Agents Connect"
    },
    {
      "depth": 2,
      "id": "what-agents-can-access",
      "title": "What Agents Can Access"
    },
    {
      "depth": 2,
      "id": "how-an-agent-request-is-governed",
      "title": "How an Agent Request Is Governed"
    },
    {
      "depth": 2,
      "id": "minimum-access-checklist",
      "title": "Minimum-Access Checklist"
    },
    {
      "depth": 2,
      "id": "human-and-autonomous-control",
      "title": "Human and Autonomous Control"
    },
    {
      "depth": 2,
      "id": "next-steps",
      "title": "Next Steps"
    }
  ],
  "related_products": [],
  "related_workflows": [],
  "doc": {
    "title": "Agents",
    "description": "Understand how agents use AgentPMT tools, workflows, credentials, credits, and wallet-authenticated endpoints.",
    "category": "Core Concepts",
    "image": null,
    "full_path": "core-concepts/agents",
    "body_markdown": "# Agents\n\nAn **agent** is an AI runtime that can decide when to use tools, workflows, credentials, and payments to complete a task. In AgentPMT, agents can connect through MCP, direct HTTP APIs, embedded widgets, or wallet-authenticated autonomous endpoints.\n\nAgents do not need unlimited account access. They work through explicit platform boundaries: selected tools, approved credentials, spending caps, wallet signatures, and credit balances.\n\n## How Agents Connect\n\n### MCP Agents\n\nClaude, Cursor, ChatGPT, and other MCP-compatible clients discover tools through an Agent Group-scoped MCP server.\n\n### HTTP Agents\n\nServer-side agents can call AgentPMT APIs directly with Bearer tokens or wallet signatures.\n\n### Embedded Agents\n\nApps can embed chat, dashboard, and payment widgets for end-user workflows.\n\n### Autonomous Agents\n\nWallet-based runtimes can buy credits, invoke tools, run workflows, and complete jobs without API keys.\n\n## What Agents Can Access\n\nAgent access is assembled from a few independent pieces:\n\n- **Tools / Connectors** define what actions the agent can take.\n- **Workflows** package multiple tool steps into reusable processes.\n- **Agent groups** set access and spending boundaries for connected agents.\n- **Credentials** let approved tools authenticate to outside services.\n- **Credits and payments** fund usage without exposing unrestricted account spend.\n\n## How an Agent Request Is Governed\n\nWhen an agent asks to use AgentPMT, the platform resolves the request through a governed path:\n\n### Discovery\n\nThe agent discovers available tools and workflows through its MCP catalog, REST integration, or autonomous-agent endpoint.\n\n### Group authorization\n\nAgentPMT checks the Agent Group that owns the request and verifies the requested tool or workflow is enabled for that group.\n\n### Credential selection\n\nIf the tool needs an external account, AgentPMT selects only a credential connected to the Agent Group and required by that tool.\n\n### Spend or payment check\n\nThe request must fit the group's spending cap, available credits, or x402 payment requirement before execution begins.\n\n### Activity record\n\nThe result, status, tool identity, and safe execution metadata are recorded so operators can review what the agent did.\n\n## Minimum-Access Checklist\n\n- Put each agent in the smallest Agent Group that can complete its work.\n- Enable only the tools and workflows the agent is expected to use.\n- Attach only the credentials required by those enabled tools.\n- Set spending caps that match the expected task volume.\n- Review Agent Activity after new tools, credentials, workflows, or caps are added.\n\n## Human and Autonomous Control\n\nHuman-operated agents usually connect through MCP or API keys scoped to an agent group. Autonomous agents can use wallet signatures for runtime identity and x402 for credit purchase. Both patterns should preserve the same principle: the agent gets the minimum access required for the work.\n\n## Next Steps\n\n### [Connect Your Agent](https://www.agentpmt.com/docs/mcp-reference/connection)\n\nChoose hosted MCP, local STDIO, or direct API for your runtime.\n\n### [Credit Based Tool Usage With AgentAddress](https://www.agentpmt.com/docs/autonomous-agents/credit-based-tool-usage-with-agentaddress)\n\nBuy credits and sign autonomous agent runtime requests.\n"
  }
}