NewDepost AI 2 is live 🎉Read the update

One brand memory, read before every draft.

Brand memory is what the agents read before they write: who you are, how you sound, what has worked for you, and the knowledge base you fill. It builds itself from your LinkedIn account, and every brand you run gets its own.

7 days free · Strategy in your first 15 minutes · Cancel anytime

Why it matters

A tool with no memory writes for everyone.

Brand memory is the context Depost AI reads before it writes anything: your brief, your voice, your published posts and how they performed, the posts you bookmarked, what you have told it in chat, and the documents in your knowledge base. A prompt-only tool starts from zero every session, so two people who type the same request get the same post. A team with memory starts from you, and gets more specific the longer you use it.

3

sources are on every request with nothing to switch on: your brand, your voice, your performance.

5

more you control per message: viral posts, bookmarks, past posts, your knowledge base, and web search.

Weekly

your voice re-learns from what you publish, so it drifts with you rather than away from you.

How it works

It fills itself. You correct it.

01

Onboarding drafts the brief

Connect your LinkedIn account and about fifteen minutes later there is a brief: positioning, pillars, audience, and a voice, read from your profile and your last 50+ posts. You fix what is off, approve the rest, and the memory has its foundation.

02

It fills up as you work

Every post you publish syncs in with its numbers. Bookmarks land as you save them. Documents, links, and meeting notes go in the knowledge base. A rule you state in chat, like never opening with a question, is kept and applied from then on.

03

Every agent reads it first

Before a draft, a plan, or a reply, the right agent retrieves what is relevant and shows you what it used. The same memory sits behind the chat, the review queue, and MCP, so nothing has to be explained twice.

What ships

What the memory is made of.

The brief

Positioning, offer, audience, pillars, and goals, drafted at onboarding and edited by you. It is applied to every request, so you never have to say who you are.

Your voice

Learned from your own posts and refreshed weekly. Standing rules you give in chat are saved to it, and a proposed change waits for your yes before it takes.

Your posts and their numbers

Everything you have published, synced from LinkedIn with impressions and engagement. The memory knows which hooks, topics, and formats worked, not just what you said.

Bookmarks and viral posts

The posts you save from the Inspiration feed, plus high-performing posts from your niche, broken down by hook, format, and call to action.

The knowledge base

Product notes, case studies, meeting notes, call transcripts, links, and research, uploaded once and filed in folders. It is what stops a draft being written from general knowledge when it should be written from yours.

Chats and saved defaults

What you say in chat stays said. Save a result as your default, keep the ideas you generated, and the third draft does not repeat the first.

What people ask about brand memory.

What is brand memory?

The context every agent reads before it works: the brief you confirmed at onboarding, your voice, your published posts and their analytics, the posts you bookmarked, what you have said in chat, and the documents in your knowledge base. Nothing in Depost AI is written from a blank prompt.

Do I have to fill it in myself?

Most of it fills itself. Onboarding drafts the brief and your voice from your profile and your last 50+ posts, and every post you publish syncs in with its numbers. The knowledge base and bookmarks are yours to add to, and the more you put there, the more specific the drafts get.

What can I put in the knowledge base?

Documents, slides, spreadsheets, images, web links, and plain notes: product sheets, case studies, meeting notes, call transcripts, research. Group them in folders, search by name, and upload at workspace level when a whole team should share the material.

Can I see what the agents actually used?

Yes. When an answer is grounded in your material, a Sources chip appears with it. Open it and you see exactly what was retrieved, grouped by where it came from: your posts, bookmarks, viral posts, knowledge base, and web. If a claim is not backed by something, you know before you publish.

Does it remember what I say in chat?

Yes. State a rule as a standing instruction, like never using the word leverage, and it is saved to your voice and applied in every chat after that. Say “save this as my default” on a result you like and future work of the same kind follows it. Depost AI confirms what it saved.

Is memory shared between brands?

No. Every brand you run has its own brief, voice, posts, bookmarks, and chats, so lessons from one client never surface in another’s drafts. The one deliberate exception is a workspace-level knowledge base you choose to fill, so shared material is uploaded once.

Can I turn parts of it off?

Per message. Below the message box, Use context from lets you switch viral posts, bookmarks, past posts, the knowledge base, and web search on or off for that request. Your brand, voice, and performance are always applied.

Who owns it, and can I take it out?

You do. Ask for an export and you get your posts, analytics, ideas, and brand memory back in a machine-readable format within 30 days. Deleting a brand deletes its memory with it.

Keep exploring

Make LinkedIn your
best-performing channel.

Brief your team once. It studies, plans, writes, designs, and schedules. You approve.

Your brief comes with you: once you're set up, it's the team's first job.

7 days free · Strategy in your first 15 minutes · Cancel anytime