Deep Research Plugin

Native OpenClaw plugin for managing multi-step research projects locally.

It gives the agent a persistent research workspace with:

• multiple research sessions tracked by id and title
• an active research context so follow-up steps can keep using the same workspace
• captured resources with URLs, summaries, notes, and source metadata
• synthesis entries by phase so findings can evolve from framing to reporting

Install

npm install
npm run build
openclaw plugins install .
openclaw plugins enable deep-research
openclaw gateway restart

Config

{
  plugins: {
    entries: {
      "deep-research": {
        enabled: true,
        config: {
          storagePath: "~/.openclaw/state/deep-research/research.json",
        },
      },
    },
  },
}

If storagePath is omitted, the plugin stores data in ~/.openclaw/state/deep-research/research.json.

Tools

• deep_research_create
• deep_research_list
• deep_research_get
• deep_research_update
• deep_research_focus
• deep_research_add_resource
• deep_research_add_synthesis

Usage

After enabling deep-research, ask for things like:

• Research the latest browser MCP support across major AI coding tools and give me a sourced comparison.
• Build a brief on EU AI Act enforcement timelines with exact dates, primary sources, and open questions.
• Compare three vendors in the passwordless auth space and separate confirmed facts from marketing claims.

The agent can create a research session, keep it active across steps, save resources as they are found, and append synthesis notes for each phase.