Cowrie Research

nansen-primary-sources-2026-04-26

4 min read

Nansen primary-source snapshot

Fetched: 2026-04-26 UTC

1) Official site positioning

Source: https://nansen.ai/

Selected signal from the site:

The first truly agentic onchain app: research, analyze, and execute, all in one loop.

Additional signals from the site:

  • Nansen no longer presents only as a dashboard or analytics subscription; the homepage emphasizes an end-to-end research-to-execution loop
  • The site claims 500M+ labelled addresses and frames Smart Money as the core discovery surface
  • The product pages visible from the homepage include Trading, Smart Money, API, Staking, Points, and Portfolio, which suggests a broad platform rather than a single analytics product

2) API docs and source-of-truth posture

Source: https://docs.nansen.ai/

Selected excerpts from the docs:

Nansen API offers programmatic access to high-quality onchain data and advanced blockchain analytics across numerous networks.

Nansen applies unique, human-readable labels to hundreds of millions of blockchain addresses.

Additional signals from the docs:

  • The docs emphasize contextualized intelligence, not just raw blockchain data
  • Official API categories include Smart Money, Profiler, Token Screener, Portfolio, Points, Hyperliquid, Agent, and Prediction Markets
  • The docs portal itself is a strong primary source because it exposes supported chains, filters, rate limits, agent access, and MCP integrations

3) AI-agent access and research-agent endpoints

Source: https://docs.nansen.ai/getting-started/agents Source: https://docs.nansen.ai/api/agent

Selected excerpts from the docs:

AI Agent Access

Your agent saves the key and handles everything else automatically.

Ask the Nansen AI agent a research question and receive a streamed answer backed by on-chain data.

Additional signals from the docs:

  • Nansen provides both a user-facing agent-auth setup flow and API-level research-agent endpoints
  • The agent API exposes separate fast and expert modes plus event-stream semantics for multi-turn research workflows
  • The published agent interface is concrete enough to treat Nansen as AI-facing infrastructure, not only as a web analytics app

4) Hosted MCP surface

Source: https://docs.nansen.ai/mcp/connecting

Selected excerpts from the docs:

The Nansen MCP server provides AI assistants with access to comprehensive blockchain analytics through a secure API.

Ensure you have sufficient credits for your usage.

Additional signals from the docs:

  • Nansen documents a hosted MCP endpoint and direct integrations for Claude Desktop, Claude Code, Cursor, and other MCP clients
  • The docs make clear that MCP usage is tied to the same API-key and credit system as the rest of the platform, which matters for how the business and control plane are structured

5) Direct execution / agentic trading

Source: https://academy.nansen.ai/articles/3306593-nansen-trading-101

Selected excerpts from the academy:

Option 1: Agentic Trading

Start a chat with Nansen AI agent (asking it to swap/trade if you specifically want to trade)

Review routes, prices, details (slippage, speed, etc).

Confirm execution

Additional signals from the academy:

  • Nansen’s own help materials describe direct execution from an AI chat flow, not just research assistance
  • Standard routing references Jupiter, OKX, and LI.FI, which reinforces that Nansen is aggregating execution paths rather than only reporting analytics
  • The mobile and web instructions both place wallet activation and direct trading inside the Nansen app surface

6) Public engineering surface

Source: https://github.com/nansen-ai

Selected signals from the public GitHub org:

Nansen AI skills for OpenClaw and Claude Code — smart money tracking, wallet profiling, token analysis, portfolio analytics, and Hyperliquid perps across 20 chains

This repo contains table definitions for certain EVM compatible blockchains.

Additional signals from the public materials:

  • The org includes ETL table-definition repos, nansen-cli, MCP packaging, and AI-skill repositories
  • GitHub reinforces that Nansen maintains public developer and data-infrastructure surfaces beyond the consumer-facing app

Notes

  • This pass did not locate a canonical standalone whitepaper or litepaper from official Nansen materials.
  • Nansen appears best cataloged as onchain analytics / labeled-wallet intelligence / agentic trading and API platform.
  • The strongest primary sources were the main site, docs portal, AI-agent and MCP docs, academy/help materials, and public GitHub organization.

Graph View