Understanding PF Create

Your AI already knows things.PF Create lets it do things.

PF Create is an MCP server that gives your existing AI assistant 190+ real-world tools. Purpose-built to cut costs — fewer tokens, cheaper images, no markups, and research that stays with you across every session.

Get started →Connection guides

50%cheaper images in batch mode

~90%fewer tokens per task

BYOKno markup on provider costs

RAGresearch auto-indexed, always available

The technology

What is MCP and why does it matter?

Think of MCP like a USB-C cable for your AI. USB-C is one universal connector that works with your laptop, phone, monitor, and hard drive — you don’t need a different cable for each device. MCP works the same way: it’s one standard connection that lets your AI plug into any tool or service that speaks the protocol. PF Create is an MCP server — one connection that gives your AI 190+ structured tools covering images, WordPress, SEO, social, memory, and files, plus 35 on-demand skills (pre-built workflows like brand-voice writing, newsletter production, and image pipelines that snap into any AI client instantly), and direct access to Gemini and Perplexity without ever leaving the AI tool you already use.

PF Create
Your AI Tool: Claude · ChatGPT Cursor · n8n
Image Generation: 20+ models · BYOK OpenAI · Gemini · FLUX
WordPress & Elementor: 95–97% fewer tokens element tree · publish
SEO & Research: 12 free tools Perplexity · citations
Social & Messaging: LinkedIn · Telegram async notifications
Memory & Files: Index once 75–94% context saving
Gemini & Perplexity: Ask any AI directly grounded · reasoning
35 On-Demand Skills: Brand voice · pipelines ~95% prompt saving

PF Create
Your AI: Claude ChatGPT
Images: 20+ models BYOK
WordPress: 95-97% fewer tokens
SEO: 12 tools Perplexity
Social: LinkedIn Telegram
Memory: RAG 75-94% less
Gemini & Perplexity: Any AI directly
35 Skills: On-demand ~95% saving

Why PF Create saves you money

Fewer tokens. Cheaper images. Research that stays.

AI API costs are driven by token volume. PF Create cuts that load at every level —

Image task token cost

Without

1200 tokens

With PF Create

120 tokens

Page editing token cost — WordPress MCP vs PF Create

WordPress MCP

12000 tokens

PF Create

400 tokens

Structured tool call token cost

Raw schema

800 tokens

PF Create

80 tokens

Agentic Memory token cost per session

Re-paste each session

2000 tokens

RAG indexed

300 tokens

Dynamic skills prompt cost

Full system prompt

900 tokens

PF Create skill

60 tokens

Perplexity research token cost

Re-run each session

500 tokens

RAG indexed

0 tokens

Before & after

What changes when you connect?

Without PF Create

With PF Create

✕Paste brand docs into every session (~5,000 tokens)

✓RAG retrieves only relevant snippets (~400 tokens)

✕Build pages by generating raw HTML — 6,000–18,000 tokens per section

✓patch_element on Elementor tree — ~90 tokens per edit, ghost preview first

✕Re-run the same research every new conversation

✓Perplexity research auto-saved — available in every future session

✕Paste a 600-token system prompt before every writing task

✓Dynamic skills invoked in ~60 tokens — no prompt pasting

✕20 separate manual sessions for product images

✓Batch 100 images at 50% off, delivered async to your library

✕Pay markup on AI costs through intermediary platforms

✓BYOK — pay providers at list price, zero markup

Without PF Create

✕Paste brand docs into every session (~5,000 tokens)

✕Build pages by generating raw HTML — 6,000–18,000 tokens per section

✕Re-run the same research every new conversation

✕Paste a 600-token system prompt before every writing task

✕20 separate manual sessions for product images

✕Pay markup on AI costs through intermediary platforms

With PF Create

✓RAG retrieves only relevant snippets (~400 tokens)

✓patch_element on Elementor tree — ~90 tokens per edit, ghost preview first

✓Perplexity research auto-saved — available in every future session

✓Dynamic skills invoked in ~60 tokens — no prompt pasting

✓Batch 100 images at 50% off, delivered async to your library

✓BYOK — pay providers at list price, zero markup

What's included

Six tool domains — each with built-in cost efficiency

Click any card to see the full capability list and cost saving for that domain.

AI research — optional & paid

Perplexity research — powerful, and saved to memory automatically

Add a Perplexity API key and your AI gains four grounded research models. Perplexity is not required — it’s pay-as-you-go on top of your PF Create plan. The key advantage: every research result is automatically saved to your RAG store and becomes available in all your future sessions, across every connected AI tool and conversation, without re-running the query.

Model	Best for	Approx. cost
Sonar	Fast grounded queries	~$0.006 / query
Sonar Pro	Multi-source synthesis	~$0.016 / query
Sonar Reasoning Pro	Analysis + chain-of-thought	~$0.016 / query
Sonar Deep Research	Exhaustive async report (20–40 min) — auto-indexed to RAG on completion	~$0.40–$1.32 / job

Reusable workflows

Reusable workflows — load once, use in every session

PF Create ships with 35+ built-in skills covering brand voice, image pipelines, and publishing workflows. You can also create your own. A dynamic skill is a markdown file you write once and save to your file store — your AI invokes it with a ~60 token call instead of pasting a 600–1,200 token system prompt every session. Skills are shared across your team on group plans and work across every connected AI tool.

System prompt pasted every session

900 tokens

Dynamic skill invocation

60 tokens

Get started

Ready to connect?

Setup takes five minutes. Note: the URL format differs by client type.

Client	MCP Server URL
Claude.ai / Desktop / Code	https://mcp.pfcreate.ca/sse
n8n / Make / Zapier / HTTP	https://mcp.pfcreate.ca

Claude.ai →Claude Desktop →ChatGPT →Cursor →All 19 clients →