Building your identity with the guided chat

Most creators have a clear "feel" they want their writing to have, but cannot put it into words. The result is AI output that sounds like everyone else, drafts that need heavy rewriting, and a brand that drifts between sessions because nothing is anchored. Build my identity surfaces the answers you already have through specific, story-shaped questions, then writes them down in a format Eden can use everywhere.

Starting the conversation

1. Open the Identity tab in the sidebar.
2. Click Build my identity.
3. A new chat opens with the guided identity-building preset active.

You don't need to prepare anything. The chat opens with a short welcome and one question. Answer in your own words. You can always come back later (the chat persists in the chat selector at the top of the workspace).

What the conversation looks like

The first turn is an educational welcome, not a question dump. It frames the session, tells you the three things you can always do (say "I don't know", attach material, leave and come back), and ends with one specific question. The question is usually about a real person you want to help, something you have personally clawed your way out of, or something you keep explaining to people.

From there the chat moves into an excavation loop:

• Mine. A sharp follow-up on whatever you just said.
• Reflect. The model quotes you back to yourself when something lands.
• Propose. It drafts a candidate mission, core idea, or hook for you to react to.
• Refine. You push back, it adjusts, you keep going.

You should feel something forming by the second or third reply, not turn ten. If an answer feels generic, the model will ask for the version you would say at 11pm to a friend instead of the version you would post online.

Things that are okay

• "I don't know" is a fine answer to anything. Most users opening this chat do not have it figured out yet, which is the point of the conversation.
• Not having raw material to drop in. You can build a useful identity from scratch by talking it through.
• Going off and coming back. The chat lives in the chat selector at the top, so you can leave at any point and pick up where you left off.
• Resisting the model's first draft of your mission. You own the result. The model offers language for you to refine, not language that gets assigned to you.

Bringing in your own writing

If you have writing or saved material that already represents you, attaching it makes the result sharper. Four ways:

• Paste a link in the chat. Articles, posts, podcasts, YouTube videos, X threads, Loom recordings, newsletters. Eden pulls what is behind the URL.
• Paste a draft or excerpt directly into the chat box. Good when the writing is not published yet or lives outside Eden.
• @ mention saved Eden items. Boards, posts, markdown docs, PDFs. Type @ and pick from the menu.
• Select items on a board open beside the chat. Selected items become attachments on your next send. Good when you have collected a stack on one topic.

What gets saved

When the model has enough material, it surfaces a Your identity, drafted card in the chat with a Save to my identity button. The fields:

• Mission. One sentence on what you are here to help people do, become, or understand.
• Point of view. Your default angle on the work.
• Core ideas. The perspectives (not topics) you want to be known for. Capture every distinct one. The field accepts up to 40, so there is no need to consolidate.
• Influences. The thinkers, writers, or creators whose lens shaped yours.
• Vocabulary. Phrases lifted verbatim from your own words. This is the field that makes generated drafts sound like you and not a tone simulator.
• Tone, rhythm, format habits. Free-text descriptors of how your writing typically reads ("blunt but not cruel", "punchy", "opens with a scene, not a thesis").
• Avoid. Things you do not want the AI to do.

Click Save to my identity and the profile lands in the Identity tab. From that moment, every Eden AI chat writes in this identity automatically.

After it is saved

Open the Identity tab and you will see:

• A status card confirming your identity is active, with a global toggle to turn it on or off.
• The full profile rendered as editable fields. Hand-edit any field at any time and click Save changes.
• A Refine in chat button (covered below).
• A history of recent versions. Click Restore on any prior version to roll back.

Your identity now affects two things:

1. Every AI chat. Generated text adopts your cadence, mission, and vocabulary by default.
2. A "My identity" chip in the chat composer. Toggle it off for a single message if you want neutral output for that reply, then back on for the next.

Updating your identity

Three paths, depending on how much you want to change:

Edit fields directly

For small tweaks ("drop that phrase from my vocabulary"), open the Identity tab and edit the fields by hand. Click Save changes.

Say "update my identity" in any chat

If you are already in a chat and want to make a change while it is on your mind, just say it:

"Update my identity. Change my mission to 'help solopreneurs trade hustle for leverage by thinking in systems.'"

or

"Add 'non-negotiable decisions' to my core ideas."

The model produces a Save to my identity card with your existing profile merged with the requested change. Accept it and your identity updates. Every untouched field is carried through unchanged, so nothing gets lost.

Refine in chat

If a lot needs to change, click Refine in chat in the Identity tab. This opens a fresh guided chat with your current identity loaded as context. The first turn asks what you want to change. If you do not know, the model offers three specific angles based on what is actually in your saved profile ("your mission is sharp on outcome but soft on who specifically, want to narrow the audience?"). You refine from there, and the model saves the merged profile when you are done.

Tips for a strong identity

A few patterns from identities that produce the best AI output:

• Pick a real person for your audience. Not "creators" or "founders". One face. Specificity in the mission cascades into every other field.
• Make every core idea pass the book test. If you could not write 50,000 words defending it, it is a take, not a core idea. Set it aside and dig deeper.
• Lift vocabulary from your own writing. The most powerful field for matching your voice is the vocabulary list. If the AI is allowed to generate phrases for you, it generates generic ones. If you give it five phrases you have actually used, it leans on them and your output sounds like yours.
• Use the Avoid list aggressively. This is where you ban the patterns that bother you. "No em dashes", "no 'in today's world'", "no guru tone". The model treats these as hard rules.

Privacy

Your identity is per-user and never shared across workspaces or accounts. Eden's analytics record only counts of fields filled (so we can answer questions like "how many users finish their identity") and never the contents of your profile.

Troubleshooting

The chat asks too many questions before proposing anything. Tell it to draft. "Take a first stab at my mission based on what you have so far." The model will produce a candidate for you to react to.

The AI's writing still does not sound like me. Two common reasons. First, your Vocabulary list is empty or thin. Add five to ten phrases you have actually used. Second, your Avoid list is empty. Add the patterns that bother you. Your voice is shaped at least as much by what you ban as by what you affirm.

I want to start completely over. Open the Identity tab, scroll to the bottom of the form, and click Clear identity. The current profile is snapshotted into history first, so you can restore it if you change your mind.

Where to go next

• For the boost actions that pair well with a saved identity, read using boosts.
• For the empty-chat presets that let you remix a board's worth of saved posts, read chatting with a board.
• For building the boards you will mine for vocabulary later, read boards as swipe files.