Big Idea

The argument in one line.

Magnific's simultaneous launch of Agents, MCP, and Flows signals that the next competitive edge in AI creative tools belongs to whoever makes their agents user-designed and context-aware across the whole team.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

You use Magnific (formerly FreePic) for AI image generation and want to know what just changed.
You work on a creative team and want AI that shares context across teammates without everyone having to touch it.
You are evaluating MCP-enabled tools and want to see what connecting a creative platform via MCP actually looks like in practice.
You build reusable production workflows and want templated automation for image and video creative work.

SKIP IF…

You are not already using or evaluating AI image and creative tools — this assumes familiarity with the space.
You want a deep technical breakdown; this is an overview walkthrough, not a hands-on tutorial.

TL;DR

The full version, fast.

Magnific just dropped three features at once: Agents that carry your knowledge base and project memory across the whole team; an MCP integration that lets you access every Magnific tool from Claude, ChatGPT, Cursor, VS Code, or Gemini without opening the Magnific UI; and Flows, which turn complex multi-step creative processes into reusable templates. The core bet is user-designed agents — you upload your references, set your preferences, and the agent learns your team's conventions rather than starting blank every session. The MCP/Agents distinction matters: MCP brings Magnific into your chat app of choice; Agents means you work inside Magnific's native interface.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 00:53

01 · Conference framing + Magnific differentiation

Host teases the visit, calls Magnific the future for AI creatives, and establishes the key differentiator: real humans who show up and want feedback.

00:53 – 01:46

02 · Triple launch announcement

Agents + MCP + Flows just launched. Magnific philosophy quote: built to stay by your side, not replace you. MCP gives agents access to everything in Magnific.

01:46 – 02:33

03 · Agents walkthrough — team context and sync

Agents reach across folders, accounts, collaborators, and spaces. Martin example: full project spun up from one plain-language request. Real-time team sync with no agent-touching required.

02:33 – 03:25

04 · User-designed agents — knowledge base and memory

No other creative agent is user-designed. Three pillars: knowledge base (PDFs/images/video), memory (session continuity), editable workflows. Never start from blank slate.

03:25 – 04:05

05 · Flows — reusable workflow templates

Build once, flow it forever. Templates curated by use case: social media, advertising, filmmaking, motion control. Chronicles of Bone movie shown as example.

04:05 – 05:12

06 · Magnific MCP — full platform from any AI client

One MCP connection at mcp.magnific.com gives access to every Magnific tool from Claude Desktop, Claude Code, ChatGPT, Cursor, Codex, VS Code, Gemini. Save workflows to Spaces from chat.

05:12 – 05:31

07 · MCP vs Agents — the key distinction

MCP: Magnific goes to you. Agents: you work inside Magnific. Same features, different entry points.

05:31 – 05:51

08 · Outro and subscribe CTA

More updates coming. Host can barely keep up with the pace of AI. Subscribe CTA. Subscribe to the mule.

Atomic Insights

Lines worth screenshotting.

Magnific's stated design principle — 'most AI agents are built to replace creatives, we built ours to stay by your side' — is the clearest positioning line in AI creative tools right now.
The difference between MCP and Agents in one line: MCP means Magnific goes to you; Agents means you go to Magnific.
Teammates who never touch an agent still get the full synced project, which removes the adoption barrier that kills most AI team workflow tools.
Flows turn one-off complex processes into run-anywhere templates, solving the rebuilt-from-scratch-every-time problem that eats creative production time.
An MCP endpoint at mcp.magnific.com gives you the full Magnific platform from Claude Desktop, ChatGPT, Cursor, Codex, VS Code, and Gemini in one connection.
The knowledge base feature means the agent carries PDFs, images, video, and text files into every conversation — no more re-briefing from a blank slate.
Agent memory is session-persistent and team-persistent: the next session picks up where you left off, and no teammate has to re-onboard the agent.
Being physically present — at a conference and inside the company's office — is still a content moat even in the AI era; this video exists because the creator showed up.

Takeaway

Three features that change how creative teams use AI.

WHAT TO LEARN

Magnific's triple launch reframes what a creative AI tool owes its users: not a smarter model, but an agent that carries your team's context, connects to your existing tools, and runs your repeatable work automatically.

User-designed agents solve the blank-slate problem: upload your references and project conventions once, and the agent carries them into every conversation from then on.
Session memory means the next conversation picks up where the last one left off — your teammates do not need to re-brief the agent either.
MCP integration lets you run the full creative platform from whichever AI chat you already live in — Claude, ChatGPT, Cursor, VS Code — without switching windows.
Flows are the automation layer: build a multi-step process once, save it as a template, and run it from any agent or chat with no rebuilding.
The MCP/Agents distinction is a useful mental model for any platform offering both: MCP means the tool comes to you; native agents mean you go to the tool.

Glossary

Terms worth knowing.

Magnific (formerly FreePic): An AI image upscaling and generation platform positioned for professional creatives. Rebranded to Magnific; still carries the FreePic legacy user base.
MCP (Model Context Protocol): An open protocol that lets AI clients like Claude or ChatGPT connect to external tools and services. Magnific's MCP endpoint makes the full platform accessible from any compatible AI chat interface.
Flows: Magnific's name for reusable workflow templates — a series of steps built once, saved to Spaces, and run by any agent or from any chat. Curated by use case and industry.
Spaces: Magnific's collaborative workspace layer — a shared environment where teams store projects, flows, and agent configurations so the whole organization can access and build on them.
Upscale Conference: A creative AI industry conference where Magnific's team presented these updates and invited creators to their San Francisco office for a firsthand preview.

Resources

Things they pointed at.

00:07productMagnific ↗

04:25linkMagnific MCP endpoint ↗

Quotables

Lines you could clip.

01:13

“Most AI agents are built to replace creatives. We built ours to stay by your side.”

Standalone positioning line, zero context needed, emotional contrast built in→ TikTok hook↗ Tweet quote

02:44

“No other creative agent is user-designed. We built ours so you could build yours.”

Tight, punchy, memorable — works as a pull-quote or cold-open→ IG reel cold open↗ Tweet quote

03:37

“Build something that works, don't rebuild it, flow it.”

Aphorism format, zero setup needed→ newsletter pull-quote↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

Magnific, formerly known as FreePic, in my opinion, the future for AI creatives. And based on my recent trip to the upscale conference, I'll show you exactly why.

But it's not because of the amazing AI updates that they dropped. It's about much more than that. So if you don't have a Magnific aka free pick subscription yet, there's a link in the description.

And a big thank you to Magnific for inviting me to your San Francisco office. What makes them really stand out is that they are actually available. Real people that communicate back.

In fact, at the upscale conference, the entire AI team was there and actually very interested in feedback and what you had to say, which now brings me back to Magnific's offices where I personally got a look at some updates that just dropped.

And what the future has in store for not only Magnific, but my opinion, the future of AI creation.

And Magnific first said this, most AI agents are built to replace creatives. We built ours to stay by your side. For the first time, what agents learn is shared across your team.

Agents plus MCP plus flows just launched.

Over MCP, the AI will have access to absolutely everything that is in Magnific.

You could see right here, it says Magnific is connecting AI to the tools you already use and making it remember you and your past decisions. And when it comes to agents, this can reach across everything in Magnific.

Your folders, accounts, usage, collaborators, and spaces.

So they understand the context around your work so they can take action with the full picture while you stay in control. So you can see right here in this post that we have Martin here. He asked the agent to spin up a full project in just one simple request.

Folders, character specs, post production, all from plain language.

Everything that Martin builds syncs in real time. Teammates who never touch the agents still get the full project and he can invite the whole team to work in the same place.

They can create new programs with spaces that can be used by the whole organization. Now you don't have to stay inside Magnific. Agents can now connect with Magnific

from Claude, ChatGPT, or from Magnific's WebUI, All made possible by the Magnific MCP.

Ask for anything, and Magnific picks the right model and brings the result back. Magnific MCP

brings in the tools we already work with. Simply put, design the agent, direct the work. So you can start with premade agents or simply design your own.

They carry your references, hold your project memory, and run inside the spaces your team already uses. And they say specifically, no other creative agent is user designed.

We built ours so you could build yours. So three ways every agent becomes yours, what it knows, what it remembers, and what it builds. We have the knowledge base.

You can upload PDFs, images, video, text files, etcetera. The agent carries them into every conversation

so you never have to start from a blank slate. The complex jobs describe the goal and the agent plans and runs every step, from first image to final result.

With access to projects and your history on Magnific, the agents can make intelligent decisions.

And when you want to go deeper, take the work into spaces, build a work flow once, save it, and run it from any agent. And when it comes to flows, check this out. Magnific states that build something that works, don't rebuild it, flow it.

Flows turn complex process into ready to run templates curated by use case and industry for you, for your team, for whatever you make next. Now, what makes it easy are gonna be these templates that we can already go with. Tons of selections that we could choose from.

All you would have to do is scroll through a few of these or even tab up at the very top what you're looking for. We have social media, advertising, filmmaking, motion control, and more like this movie, the chronicles of bone.

Again, all built with everything I'm showing you today. And when it comes to Magnific MCP, again, everything is sort of intertwined.

We get the full Magnific platform from any AI client. Connect one, access every Magnific tool from the chat that you already use. It gives us the exact instructions on how to connect this from claw desktop, clawed code, chat GPT, cursor codex, v s code, Gemini.

You can even make a workflow from the chat. Right here, describe what you want it to do, save it as a workflow in spaces, and run it from any chat and share it with the team when ready. Now these pages really do share some very good input, so make sure you check the links in the description.

I'll be including every link that we're looking at today. So right here, you can see the difference between the two. MCP and agents.

MCP Magnific goes to you. So you're using it in the specific app of your choice like chat GPT agents.

You work inside of Magnific. So I know that this is a lot to take especially if you're more on the beginner side, because so much actually was dropped at once. And that doesn't even include other major updates that just dropped, and a lot more game changing updates also coming very soon.

Honestly, we're just warming up. So to get the latest updates, AI tips, and tricks, make sure you hit that subscribe button because I'll be sharing as much as I possibly can, if in fact, can actually keep up with the speed of AI. Until then, I'll see you next time.

Subscribe to the mule. If it's the

The Hook

The bait, then the rug-pull.

The host walked into the Upscale Conference and Magnific's San Francisco office and came back with a video that buries the lede on purpose. The title promises big updates — the first card on screen says it is about much more than that.

Frameworks