AgentLed CLI

The automation engine built for AI agents. Give any MCP client workflows, persistent memory, and 100+ integrations — one API key.

Prerequisites

Node.js 18 or newer — check with node --version
npm or npx — ships with Node.js
A browser — only required for the interactive setup flow. Skip to Headless install if you're running in CI or from an AI agent without browser access.

Quick Start (interactive)

One command: opens a browser, creates an account, issues an API key, and configures your MCP client (Claude Code, Codex, Cursor, …).

# Run without installing
npx @agentled/cli setup

# Or install globally first
npm install -g @agentled/cli
agentled setup

Verify the install

agentled --version
agentled --help

Headless install (AI agents & CI)

agentled setup opens a browser — not an option for AI agents, CI runners, or remote shells. For those, skip setup entirely and invoke the MCP server directly with an API key.

Sign in at agentled.app, go to Workspace Settings → Developer, and create an API key. It looks like wsk_....
Export it in your environment (or pass it inline):

export AGENTLED_API_KEY=wsk_...

# Run the MCP server (what Claude Code / Codex connects to)
npx -y @agentled/mcp-server

To register the MCP server with an agent CLI in one shot:

# Claude Code
claude mcp add agentled -e AGENTLED_API_KEY=wsk_... -- npx -y @agentled/mcp-server

# OpenAI Codex
codex mcp add agentled -e AGENTLED_API_KEY=wsk_... -- npx -y @agentled/mcp-server

See the Developers page for full MCP server docs.

What is AgentLed?

AgentLed is an automation engine designed for AI agents. Unlike traditional workflow tools (n8n, Zapier, Make) that require humans to build and maintain every step, AgentLed lets AI agents create, execute, and improve workflows autonomously.

What makes it different:

1.Business-level memory — Knowledge Graph stores company context, ICP definitions, scoring models, and workflow outcomes. Shared across all workflows and users. Every run compounds on the last.
2.User-level context — Individual preferences, conversation history, and personal workflow patterns. Each user gets a personalized automation experience within the shared business context.
3.100+ integrations, one credit system — LinkedIn, Hunter, CRMs, AI models, web scraping, video generation — all through unified credits. No API key juggling.
4.MCP-native — Works from Claude Code, Codex, Cursor, or any MCP client. Your AI assistant becomes a workflow operator.

Available MCP Tools

Category	Tools
Workflows	create, update, add/remove steps, validate, publish, export, import
Executions	start, stop, retry, list, get execution details + timeline
Knowledge Graph	query edges, get rows by list, get rows by IDs, get text, scoring history, list knowledge bases
Testing	test AI actions, app actions, code actions
Snapshots	create, restore, delete, list snapshots + drafts
Chat	conversational workflow building with context
Workspace	get workspace info, branding, app actions, integrations list

Example Workflows

Lead Enrichment & Outreach

agentled create "Find fintech CTOs in Europe, enrich via LinkedIn + Hunter,
score by ICP fit, draft personalized outreach, save to Knowledge Graph"

VC Investor Matching

agentled create "Match this startup against our 2,000+ investor database.
Score by sector focus, stage preference, check size. Compare with last round outcomes."

Content Production

agentled create "Scrape trending topics in our niche, generate 5 LinkedIn posts,
create thumbnail images, schedule publishing for the week"

Credit System

One API key. All integrations run on unified credits.

Capability	Credits	Without AgentLed
LinkedIn enrichment	50	LinkedIn API ($99/mo+)
Email finding (Hunter)	5	Hunter.io ($49/mo)
AI analysis (Claude/GPT)	10–30	Multiple API keys
Web scraping	3–10	Apify ($49/mo+)
Image generation	30	DALL-E / Midjourney
Video generation (8s)	300	RunwayML ($15/mo+)
Knowledge Graph storage	1–2	Custom infrastructure

Two Layers of Memory

Other tools start from zero every run. AgentLed maintains two persistent memory layers:

Business-Level Memory

Company context shared across all workflows and users. ICP definitions, scoring models, client profiles, workflow outcomes, learned patterns. One workflow learns it — every workflow benefits.

Example: investor scoring accuracy improving from 62% → 89% over 12 runs with zero manual tuning.

User-Level Context

Individual preferences, personal workflow history, and conversation patterns. Each team member gets personalized automation within the shared business context. Your agent remembers how you work.

Example: your agent learns you prefer concise reports on Mondays and detailed analysis on Fridays.

Pricing

Pro

€24–45/mo

For individuals and small teams getting started with AI workflows.

Teams

€87–118/mo

For teams running production workflows with Knowledge Graph and collaboration.

Enterprise

Custom

White-label, on-premise, dedicated support, custom integrations.