AgentKey Integration: Web, Social, and On-Chain Data Access for AI Agents

AgentKey connects your AI agent (Claude Code, Claude Desktop, Cursor, Codex, Gemini CLI, Manus, and 40+ others) to external data sources including web search, social media, on-chain crypto data, finance, e-commerce, and business information through a single subscription. Once installed, your agent can browse Twitter, search LinkedIn, scrape any webpage, read YouTube transcripts, pull token prices, and summarize on-chain wallet activity without needing individual API keys for each provider.

What It Does

AgentKey provides your AI agent with the following capabilities, all routed through a single cloud-side integration:

• Web search: Search engines, news, and real-time discovery across the open web
• Web scraping: Clean article extraction, structured data, and metadata from any URL
• On-chain / Crypto: Token prices, market data, on-chain activity, and project metadata
• Social & content: Public posts, videos, and discussions across major social and content platforms
• Finance: Stock and FX quotes, technical indicators, company financials, and earnings
• E-commerce: Product listings, prices, reviews, and best-sellers across major marketplaces
• Business: Company, funding, investor, and people data for market and competitive research

Underlying providers are routed automatically and grow over time. Your agent can call find_tools to see what is currently available. The sources also list Maps and Weather as planned capabilities.

According to the official documentation, before AgentKey, 10 tasks required 10 API keys and 10 separate bills. With AgentKey installed, your agent gains all these data capabilities automatically with no per-provider registrations and no juggling separate API bills. One subscription covers everything.

Concrete use cases from the sources include:

• Asking your agent what someone has been saying on Twitter lately: without AgentKey, tweets are blocked; with AgentKey, the agent pulls all relevant tweets and summarizes them.
• Asking about product sentiment on Instagram: without AgentKey, it is blocked and requires login; with AgentKey, the agent scrapes real posts and organizes them by sentiment.
• Asking what a YouTube or Bilibili video covers: without AgentKey, the agent cannot read it because there are no subtitles; with AgentKey, the agent reads the video transcript and extracts key points.
• Finding Reddit threads about a specific pain point: without AgentKey, the request returns a 403 error; with AgentKey, the agent finds relevant threads and extracts solutions.
• Checking a competitor or candidate on LinkedIn: without AgentKey, there are 403 access issues; with AgentKey, the agent opens the page and summarizes key info.
• Asking what is trending on Douyin or TikTok: without AgentKey, the agent cannot scrape the hot list; with AgentKey, the agent pulls trending topics and tags.
• Asking what a webpage says: without AgentKey, the agent returns raw HTML; with AgentKey, the agent extracts the content and explains it clearly.
• Asking what a GitHub repo does: without AgentKey, you have to click through yourself; with AgentKey, the agent reads the README and Issues and gives a one-line summary.
• Asking what a wallet or fund has been buying lately: without AgentKey, you have to click through a block explorer; with AgentKey, the agent summarizes recent transactions and positions.

Setup

One-Command Install (Recommended)

The installer auto-detects every agent on your machine (40+ supported agents from the vercel-labs/skills supported-agents list, including Claude Code, Codex, Gemini CLI, and Cursor CLI) and configures each one. A browser tab opens for login, then you are done.

macOS / Linux

curl -fsSL https://agentkey.app/install.sh | bash

Windows (PowerShell)

irm https://agentkey.app/install.ps1 | iex

After installation, restart your agent. Then ask it something that needs the internet, for example: "What has Musk been tweeting about lately?" No API key to copy, no JSON to edit.

Advanced Install Options (CI / Specific Agents / Manual Two-Step)

The installer auto-detects which AI agents you have on this machine by probing well-known config directories and binaries from the vercel-labs/skills supported-agents list and pre-selects them with no multi-select prompt. You can override this with the following flags:

# Non-interactive (CI / unattended): install to every detected agent, no prompts
curl -fsSL https://agentkey.app/install.sh | bash -s -- --yes

# See which agents the installer would auto-select on this host (and exit)
curl -fsSL https://agentkey.app/install.sh | bash -s -- --list-agents

# Only install the skill for specific agents (overrides auto-detection)
curl -fsSL https://agentkey.app/install.sh | bash -s -- --only claude-code,cursor

# Skip our agent detection; let `skills` CLI install for every agent it finds
curl -fsSL https://agentkey.app/install.sh | bash -s -- --all-agents

# Only the skill, or only the MCP auth
curl -fsSL https://agentkey.app/install.sh | bash -s -- --skip-mcp
curl -fsSL https://agentkey.app/install.sh | bash -s -- --skip-skill

PowerShell equivalents: -Yes, -ListAgents, -Only, -AllAgents, -SkipMcp, -SkipSkill.

Manual Two-Step Install

If you prefer to run the two underlying commands yourself, or the one-line installer cannot reach your machine, use this approach:

# 1. Install the skill into every detected agent
npx skills add chainbase-labs/agentkey

# 2. Authenticate and register the MCP server
npx -y @agentkey/cli --auth-login

Headless / SSH / Docker

The auth step always tries to open a browser and prints the URL. On a machine without a usable display, copy the printed URL to any device with a browser to finish authentication. If you prefer typing the key manually, run npx -y @agentkey/cli --setup which opens an interactive wizard instead.

Manual Config for Unsupported Agents

MCP auto-configuration covers Claude Code, Claude Desktop, and Cursor. For Codex, OpenCode, Gemini CLI, Hermes, Manus, or Linux Claude Desktop, the skill still installs automatically but you need to paste this MCP snippet into the agent's own config (the path varies per agent):

{
  "mcpServers": {
    "agentkey": {
      "type": "http",
      "url": "https://api.agentkey.app/v1/mcp",
      "headers": { "Authorization": "Bearer ak_..." }
    }
  }
}

Then restart the agent. The skill's first-run activation will also walk you through this.

Claude Code Plugin Mode

Install straight from the marketplace. The plugin prompts you for your AgentKey API key on enable and wires the MCP server for you, so there is no second @agentkey/cli step:

# Public install
claude plugin marketplace add chainbase-labs/agentkey
claude plugin install agentkey@agentkey

# Or from a local checkout, for development
claude plugin marketplace add /absolute/path/to/agentkey
claude plugin install agentkey@agentkey

On enable, Claude Code prompts for AGENTKEY_API_KEY (stored in your OS keychain) and injects it into the plugin's .mcp.json via ${user_config.AGENTKEY_API_KEY}. Reload a local checkout with claude plugin update agentkey after edits. The sources note that day-to-day skill iteration is still fastest via the skills-CLI path; the plugin path is the one-step option for Claude Code users.

Codex Plugin Mode

Install from the marketplace bundled in this repo. Auth is OAuth (browser sign-in on install), so there is no API key to paste and no second @agentkey/cli step:

# Public install
codex plugin marketplace add chainbase-labs/agentkey
# then run /plugins inside Codex and install AgentKey,
# or: codex plugin install agentkey@agentkey

The plugin manifest lives in .codex-plugin/plugin.json; it bundles the same skill plus a remote-HTTP MCP entry (.codex-plugin/mcp.json) that authenticates against https://api.agentkey.app/v1/mcp via MCP OAuth (RFC 9728 discovery). Sign in with your AgentKey account when Codex prompts you.

Self-Host or Develop Against AgentKey

Install from a local checkout:

git clone https://github.com/chainbase-labs/agentkey.git
cd agentkey

# 1. Install your working tree into every detected agent
npx skills add .

# 2. Register the MCP server (if you haven't already)
npx -y @agentkey/cli --auth-login

npx skills add . accepts a local path or a file:// URL. Run it again after each edit to skills/agentkey/SKILL.md. The MCP step only needs to run once per machine.

If you are iterating on the MCP server itself, the server lives at AgentKey-Server/ (Go) and exposes the MCP endpoint at /v1/mcp. Run a local server with make run and point your MCP config at http://localhost:8081/v1/mcp to test changes end-to-end.

Uninstall

One command cleans every agent and config file.

macOS / Linux

curl -fsSL https://agentkey.app/uninstall.sh | bash

Windows (PowerShell)

irm https://agentkey.app/uninstall.ps1 | iex

This removes the skill from every agent, strips the agentkey MCP entry and API key from all MCP client configs, and clears caches and logs. Pass --keep-marketplace (bash) or -KeepMarketplace (PowerShell) to retain the Claude Code plugin marketplace entry.

If you prefer a manual two-step uninstall:

# 1. Remove the skill from every agent
npx skills remove chainbase-labs/agentkey

# 2. Delete the "agentkey" entry under mcpServers in each MCP client config:
#    - Claude Code:     ~/.claude.json
#    - Claude Desktop:  ~/Library/Application Support/Claude/claude_desktop_config.json  (macOS)
#                       %APPDATA%\Claude\claude_desktop_config.json                      (Windows)
#    - Cursor:          ~/.cursor/mcp.json

The one-command uninstaller additionally cleans npm/npx caches, legacy shell rc entries, CLAUDE.md sections, and MCP stdio logs. The sources recommend using the one-command uninstaller if you want a fully clean slate.