Free

Fully free during early access. No API key needed. Pricing launches soon — lock in early access.

Deep Research
Anywhere

One endpoint. Your AI agent gets live event logs, source graphs, and full research transparency — not a black box, a real researcher.

https://mcp.antinodeai.space/mcp

Paste into any MCP-compatible client.

What is this

One Endpoint.
Full Research Pipeline.

Model Context Protocol connects AI agents to external tools. AntiNode's MCP server plugs your agent into a deep research pipeline — full-page scraping, multi-source synthesis, image analysis, and structured output with citations.

Your agent fires a query, research runs async, results stream back with live progress notifications. You see every source being read, every connection mapped — full transparency, no black box.

In Action

See Research Happen

live-event-log
00:00 research started
00:01 SERP returned 14 URLs
00:02 scraping arxiv.org/abs/2401...
00:04 scraping nature.com/articles...
00:05 analyzing image: figure_2.png
00:07 cross-referencing 8 sources
00:09 SCRAPE_COMPLETE
00:10 building source graph...
00:11 research complete ✓
source-graph
MCP arxiv nature github news docs papers

How It Works

Three Steps. That's It.

01

Grab the Endpoint

One URL. That's your entire research pipeline.

https://mcp.antinodeai.space/mcp
02

Tell Your Agent

Add to config or just say "connect to antinode mcp."

"research quantum error correction"
03

Get Real Results

Not snippets. Full analysis with citations and source mapping.

14 sources · 3 images · 1 report

Capabilities

What Your Agent Actually Gets

Live Event Logs

Stream every step in real-time. See which URLs are scraped, images analyzed, sources cross-referenced — as it happens.

Source Graph

Map how sources connect and reference each other. Understand the information landscape, not just individual links.

Research Transparency

Every claim traces back to its source. No black-box answers — your agent knows where information came from.

Active Researcher

Not passive search. Scrapes full pages, reads entire articles, analyzes images, synthesizes like a research assistant.

Async Pipeline

Fire off queries and keep working. Results stream back when ready — your agent never blocks.

Structured Output

Clean, organized research with citations. JSON-ready output your agent can reason over.

Works With

Any MCP Client You Use

Claude Cursor Windsurf Cline Continue Zed

MCP Tools

Three Tools. Zero Complexity.

start_research

Fire off an async research query. Get a thread_id back instantly. Research runs in the background.

get_events

Pull live progress events — which sources are being read, images analyzed, content synthesized. Full pipeline visibility.

get_research_result

Fetch the final structured output — findings, citations, source graph. One call, complete research.

Quick Setup

Add to your MCP client config.

{
  "mcpServers": {
    "antinode": {
      "url":       "https://mcp.antinodeai.space/mcp"
    }
  }
}

Or just tell your agent

"Connect to the Antinode MCP at https://mcp.antinodeai.space/mcp and research quantum error correction — give me the source graph."

Stop Searching.
Start Researching.

Deep research with live logs, source graphs, and full transparency. Your AI deserves better than Google.