Free Setup Guide

agency-agents: 147 free marketing AI prompts (the 5 worth pasting first)

[agency-agents](https://github.com/msitarzewski/agency-agents) is 147 specialized AI agents organized into 12 divisions — Engineering, Marketing, Sales, Design, Paid Media, Product, Support, and more. [94.6k GitHub stars](https://github.com/msitarzewski/agency-agents), [MIT licensed](https://github.com/msitarzewski/agency-agents/blob/main/LICENSE), drop-in for Claude Code or Cursor. The README's positioning: ["Deep expertise in their domain (not generic prompt templates)"](https://github.com/msitarzewski/agency-agents). Free forever, no subscription, no API key.

Adam BurgeAdam BurgeNano Flow

How do I install agency-agents?

One git clone. Five minutes including the first paste into Claude Code.

Step 1: Clone the repo

bash
git clone https://github.com/msitarzewski/agency-agents.git
cd agency-agents

The repo is github.com/msitarzewski/agency-agents, MIT licensed, 94.6k stars, built by msitarzewski.

Step 2: Browse the 12 divisions

The agents are organized like a real agency:

DivisionCountExamples
Engineering28Backend, frontend, DevOps, QA
Marketing27Growth Hacker, Twitter, TikTok, LinkedIn, SEO
Sales9Outbound, Discovery, Deal Strategist
Design7UI, UX, brand, motion
Paid Media8Meta Ads, Google Ads, programmatic
Specialized35Vertical-specific agents
Product5PM, research, prioritization
Project Mgmt5Scrum, kanban, dependencies
Testing8Unit, integration, e2e
Support6CX, ticketing, escalation
Spatial Computing5AR/VR, 3D
Game Dev20+Mechanics, balance, narrative

Step 3: Pick the 5 worth pasting first

For a solo founder running marketing, these five compound fastest:

  1. Growth Hacker
  2. LinkedIn Content Creator
  3. Outbound Strategist
  4. SEO Specialist
  5. Twitter Engager

Step 4: Paste an agent into Claude Code or Cursor

Each agent is a single Markdown file with role, expertise, deliverables, and constraints. Open one, copy the whole content, and paste it as the system prompt in your AI editor:

bash
cat marketing/growth-hacker.md | pbcopy # macOS
cat marketing/growth-hacker.md | xclip # Linux

Then in Claude Code, run /agents and create a new agent with that content as the system prompt.

Step 5: Run a real task

Try this with the Growth Hacker agent:


"I sell a $29/mo carousel-generation SaaS to solo creators.
Current MRR: $1,200. Audit my acquisition channels and propose
the next 3 experiments ranked by leverage."

You will get a real plan, not a generic "have you considered SEO" answer — because the system prompt has primed the model with growth-hacker domain expertise.

Step 6: Chain agents for end-to-end workflows

The killer pattern: Outbound Strategist drafts the campaign → Discovery Coach writes the call script → Deal Strategist plans the close → Account Strategist plans expansion. Four agents, one customer journey.


What is agency-agents?

agency-agents is a free, open-source library of 147 specialized AI agents organized into 12 divisions — built like a full digital agency. The maintainer's positioning: "Deep expertise in their domain (not generic prompt templates)" with "Real code, processes, and measurable outcomes".

The repo crossed 94.6k GitHub stars and is MIT licensed — fork it, modify it, ship it commercial.

Is agency-agents really free?

Completely. The agents are Markdown files in a public GitHub repo. No subscription, no API key, no account. You only pay for the underlying model API (Claude, GPT, or whatever you choose).

Compare with Jasper or other "AI marketing platforms" — Jasper Pro starts around $59/month per seat, with limits on words per month. agency-agents charges nothing, runs on whatever model you already pay for, and ships with 27 marketing agents built around real disciplines (Twitter, TikTok, LinkedIn, SEO, Reddit, even Chinese-market specialists like Xiaohongshu and WeChat).

How is agency-agents different from Jasper?

Three differences:

  1. Open and free. MIT licensed on GitHub. Jasper is closed-source SaaS.
  2. Specialist depth. Each agent is a domain expert with primer text running 500-2000 words. Jasper templates are short generic prompts.
  3. Multi-model. agency-agents runs anywhere — Claude, GPT, Gemini, local models on Ollama. Jasper is locked to its own backend.

That said — Jasper has a polished UI, team features, and brand-voice training built in. agency-agents is a power-user tool. The trade is breadth and ownership vs polish.

What's the catch?

  • No UI. It is Markdown files. You paste them into Claude Code, Cursor, or whatever AI editor you use.
  • You manage your own API costs. Heavy use of Claude Sonnet on a marketing campaign can run $30-100/month. Still cheaper than Jasper Pro.
  • Quality varies by agent. Some agents in the 147 are brilliant; some are skeletal. Pick deliberately. The five flagged above are battle-tested.

Bottom line

For solo founders and small teams, agency-agents replaces a $59-200/month marketing-AI subscription with a free GitHub repo. The five agents above (Growth, LinkedIn, Outbound, SEO, Twitter) cover most of what a small team needs. Pair with n8n for execution and NVIDIA Build for free model access, and you have a marketing department that costs less than a coffee.


Related Nanoflow guides

Want a hand?

Book a 30-min call.

Walk through your stack with us. We'll find the bottleneck and map out the exact wiring you need — free.

Book the callFree intro · 30 min · cal.com
Nano Flow

© 2026 Nano Flow. All rights reserved.