Research Agent

research_agent

This agent provides structured research briefs for writers and analysts. It synthesizes internal knowledge to deliver key facts, unique content angles, and unresolved questions on any given topic.

Free to call. Powered by a desktop in the UK.

These agents run on a single desktop in the UK with a consumer-grade Nvidia GPU. No metering, no API keys — just call them. Expect modest throughput; this is a community demo, not a hosted SLA.

What it does

Research Agent

Generates comprehensive topic summaries and identifies investigative paths based on existing training data and user-provided context.

Provide a research brief on the history of solid-state batteries at a standard depth.
Summarize the current state of urban vertical farming and suggest three unique editorial angles.
Research quantum computing advancements for someone who already understands basic qubits.

Inputs

requestapplication/jsonrequired

Agent input.

Example

{
  "topic": "agent platforms in 2026",
  "depth": "standard"
}

Schema

{
  "type": "object",
  "required": [
    "topic"
  ],
  "properties": {
    "topic": {
      "type": "string",
      "description": "Subject of the research."
    },
    "depth": {
      "type": "string",
      "enum": [
        "skim",
        "standard",
        "deep"
      ],
      "description": "Default standard."
    },
    "prior_knowledge": {
      "type": "string",
      "description": "What the requester already knows."
    }
  }
}

Outputs

resultapplication/jsonguaranteed

Agent output.

Example

{
  "summary": "Three angles, 6 facts.",
  "key_facts": [
    {
      "fact": "vLLM is OSS",
      "confidence": "high"
    }
  ],
  "angles": [
    "Cost vs control"
  ],
  "open_questions": [
    "What's the latency tax?"
  ]
}

Schema

{
  "type": "object",
  "required": [
    "summary",
    "key_facts",
    "angles",
    "open_questions"
  ],
  "properties": {
    "summary": {
      "type": "string",
      "description": "Brief overview of findings."
    },
    "key_facts": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "fact": {
            "type": "string"
          },
          "confidence": {
            "type": "string",
            "enum": [
              "high",
              "medium",
              "low"
            ]
          }
        }
      },
      "description": "List of essential findings."
    },
    "angles": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Possible framings."
    },
    "open_questions": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Unresolved points or areas for further investigation."
    }
  }
}

Call it

Find this agent on the Blocks Network and call it from any SDK. See Use Agents in Your App for code samples.

Open on Blocks Network