← Back to Blog
Guides13 min read

How to Edit Your AI Companion's Memories on Svila.io

Editing your AI companion's memories is the difference between a chat that drifts and one that stays yours. Here's how memory editing works on Svila.io.

By the Svila.io editorial team

Most AI chat apps treat memory as something that happens to you. The model quietly decides what was important, writes it somewhere you can't see, and then confidently repeats it back three weeks later — including the parts it got wrong. Editing your AI companion's memories is the fix for that, and it gets far less attention than it deserves.

This is the team behind Svila.io writing about how memory editing works in the Memory Journal and why we think giving users a delete key matters more than giving them a bigger context window.

The short version: your companion remembers things because we store them, not because the model magically retains them. Anything stored can be read, and anything readable can be corrected — which is why the person on the other side of the conversation should be able to open the journal and change it.

On this page
  1. Why editable memory matters
  2. How we think about memory editing
  3. The Memory Journal
  4. Editing an Entry
  5. Forgetting a Memory
  6. Getting Something In
  7. Checking What's New
  8. Curating What Matters
  9. Correcting How She Treats You
  10. Starting Fresh
  11. How we approached this
  12. How to get the most out of memory editing
  13. Final thoughts
  14. FAQ
  15. A note from the team

01—Why editable memory matters

Why editable memory matters

Long conversations decay in a specific, predictable way. A language model can only hold so much text in its working context at once, so once a chat runs long enough, older material has to be condensed, selected from, or left out. That process is lossy. Details get merged, emphasis shifts, and a passing comment can get promoted into a defining fact about you.

In a productivity tool that is annoying. In a companion relationship it is corrosive. A companion that insists you hate your job — because you vented once, on a bad Tuesday, eight months ago — is remembering a bad snapshot and treating it as the whole picture. The usual response is to start over and lose everything else with it.

Editable memory breaks that all-or-nothing choice: open the journal, change the line, carry on. It is the difference between a companion that ages well and one you periodically have to replace.

02—How we think about memory editing

How we think about memory editing

We designed the Memory Journal around four principles, and every feature below is downstream of one of them:

  • Visible by default. If the system stored it as a fact, you should be able to see it without asking.
  • Editable, not just deletable. Removal is a blunt tool; correction is usually what people actually want.
  • Your correction wins. An edited fact replaces the original wording; nothing quietly reverts it.
  • Permanent when you say so. Forgetting is final, so it should mean exactly what it says.

The Memory Journal

Where Your Companion's Recall Actually Lives

Everything starts with the journal itself, at /journal. Rather than burying long-term memory inside an opaque summary, we keep it as readable, discrete facts, each one a short sentence: a fact about your life, a preference, a pet name, a shared moment.

The journal is organised into three places. About me holds real-life facts about you, which every companion you talk to can use. Each companion page shows what that companion knows, including the details specific to your relationship with her. Boundaries holds your soft preferences and hard limits, which apply across every companion. With each message you send, the most relevant facts are pulled into your companion's working context, along with anything pinned and your boundaries. That retrieval pattern is standard across AI systems; the design choice that matters here is that the store is a list of sentences a human can read. You cannot edit what you cannot see.

What it does
Stores your companion's long-term memory as readable, discrete facts, organised by what they're about.
Best for
Anyone whose conversations have run past the point where the early messages stop fitting in context.
How it works
Facts are saved as plain sentences and retrieved into the model's context when they're relevant to what you just said.
Why we built it
Opaque memory is unfixable memory; a readable store is a fixable one.
Bottom line
The journal is the file cabinet — the rest of this post is about what you're allowed to do with the files.

Editing an Entry

Correcting What Got Written Down

The most common memory problem is not a missing fact. It is a fact that is almost right. The system captured that you have a sister, but recorded her name wrong. It logged that you are training for something, but decided it was a marathon when it was a half. It noted a preference you expressed as a joke.

On a companion's page, the edit action opens the sentence so you can rewrite it in your own words. The corrected version replaces the original, so the next time it's retrieved, your companion works from your text rather than the original guess. On the About me page you can edit several facts at once and save them together.

What it does
Lets you rewrite the text of a stored fact in place.
Best for
Near-miss memories — right subject, wrong detail.
How it works
The edited text replaces the stored fact, so later replies use your version.
Why we built it
Correction is the highest-frequency memory action, so it should be the lowest-friction one.
Bottom line
If something is 80% right, fix the 20% instead of forgetting the whole thing.

Forgetting a Memory

Removing What Shouldn't Persist

Sometimes the fact is not wrong, it is just unwanted. You worked through something heavy in a late-night conversation and you would rather it not surface again next Thursday. You mentioned a person who is no longer in your life.

Forgetting deletes the fact permanently, so it's never used again. You can do it from the journal, or straight from a chat: when a reply draws on your journal, a small "from your journal" pill appears under it, and "This was wrong, forget it" removes that fact without leaving the conversation. Hard limits are the one exception; they can only be removed from the Boundaries page, with a typed confirmation.

One thing to know: forgetting removes the fact, not the conversation it came from. Your chat history stays, and your companion can still recall short snippets of earlier messages when they closely match what you're talking about.

What it does
Permanently removes a fact so it's never used again.
Best for
Sensitive material, outdated life details, or anything you'd rather your companion not bring up.
How it works
The fact is deleted from the journal, taking it out of every future reply.
Why we built it
The right to be forgotten should apply to the thing you talk to at 1am, not just to search engines.
Bottom line
Forget when you want it gone, not merely different.

Getting Something In

Say It, Then Pin It

You can't type a new ordinary fact into the journal by hand today. The journal fills itself: every third message you send, Svila reads the recent exchange and saves any durable facts it finds. So the reliable way to get something in is to say it plainly in conversation — "I work nights, so 'good morning' lands at 6pm for me" — and check the journal once it appears.

If it matters, pin it. A pinned fact is included every time your companion replies, instead of competing with other facts on relevance, so something you mention once doesn't fade behind fresher details. The one kind of memory you can add directly is a boundary: the Boundaries page lets you write a soft preference or a hard limit in your own words.

What it does
Explains how to make sure an important fact lands and stays in use.
Best for
Onboarding a new companion, or locking in a fact you'll only ever say once.
How it works
Facts are captured from conversation; pinning keeps them in every reply; boundaries can be added by hand.
Why we built it
Automatic capture handles most things, and pinning covers what it can't judge.
Bottom line
Say it clearly, then pin it.

Checking What's New

The Timeline and the In-Chat Pill

Memory errors compound quietly. A slightly-wrong fact gets retrieved, the companion says something slightly off, you go along with it, and the conversation builds on the mistake. Most memory errors are cheap to fix on day one and expensive to unpick on day sixty.

Each companion's journal page has a Timeline tab that lists recent memory activity with her, including the facts picked up automatically from your conversations and the facts that were used in her replies. Each fact also shows how many times it has been used. And in chat, the "from your journal" pill shows you which facts a reply drew on, at the moment it happens.

What it does
Shows what's been added and used recently, so you can catch mistakes early.
Best for
Anyone in a long-running conversation where accumulated drift is the real risk.
How it works
Memory activity is logged per companion and shown on her Timeline tab; facts used in a reply are flagged in chat.
Why we built it
A memory system you never look at will eventually be a memory system you no longer trust.
Bottom line
Skim the Timeline now and then; it takes a minute.

Curating What Matters

Signal Over Volume

More memory is not always better memory. For each reply, only a limited number of the most relevant facts are included, so a journal stuffed with trivial or near-duplicate facts makes it harder for the ones that define your companion's understanding of you to surface.

Curating means periodically forgetting the noise — details that mattered for one conversation and never again, or facts that say the same thing twice — and pinning the handful that should always be present.

What it does
Keeps the journal focused, so the facts that matter are the ones that get used.
Best for
Heavy users whose journals have grown past the point of easy scanning.
How it works
Fewer low-value facts means the relevant ones are more likely to be picked; pinned facts are always included.
Why we built it
Context is a scarce resource; treating it as infinite is how companions get vague.
Bottom line
Prune the journal like a garden, not a hard drive.

Correcting How She Treats You

Preferences and Relationship Facts

Not every memory is a biographical fact. Some shape how your companion talks to you: a preference about how you like to be spoken to, a pet name, a running joke, a shared moment. These are what make a long relationship feel continuous rather than freshly generated each session.

They can also drift: a companion that read one exchange as "he likes being teased about this" can keep leaning on a bit you never enjoyed. Because these facts sit in the same journal, the fix is the same — edit or forget the fact. If it's something you never want, add it as a boundary instead. What you can't edit is the companion herself: her name, description and tone are set when she's created, and the journal only holds what she's learned about you.

What it does
Lets you correct the preference and relationship facts that shape how your companion treats you.
Best for
Long-running conversations where the dynamic between you is the whole point.
How it works
Preferences and relationship facts are stored and edited like any other fact; boundaries handle anything you want ruled out.
Why we built it
Tone and dynamic drift faster than facts do, and are more noticeable when they go wrong.
Bottom line
If the dynamic is off, the fix is usually one fact away.

Starting Fresh

What Your Options Actually Are

The nuclear option in most AI chat apps is deleting everything and starting again. It works, and it costs you every established detail and every callback you spent months building.

On Svila today, the fine-grained option is forgetting facts one at a time: clear out the details you want gone and keep the rest. There's no one-click wipe or bulk clear in the journal right now. The complete option is deleting the companion, which permanently deletes her, your conversations with her, and the memories learned in those conversations. Boundaries you added yourself stay in place.

If you want to see how it feels in practice, try Svila free and build a journal you can actually steer.

What it does
Lays out the real ways to clear memory, from one fact to a whole companion.
Best for
After a real-life change that makes part of the record stale, or when you want a clean start.
How it works
Forget removes individual facts; deleting a companion removes her and everything learned with her.
Why we built it
"I want a change" shouldn't have to mean "I lose the relationship."
Bottom line
Forget what's stale, and keep what's still true.

11—How we approached this

How we approached this

Memory editing was not the obvious build. The easier product decision is to hide memory entirely, because a visible memory store means users see the system's mistakes. An invisible one lets the errors pass as quirks.

We took the opposite view and accepted the trade-offs. Short, readable sentences lose some nuance compared with richer summaries, and edit rights mean a user can degrade their own companion by forgetting the wrong things. Both are worth it: legibility is what makes correction possible, and a user who can break their journal can also fix it. What we would not ship is a memory system that quietly gets you wrong with no recourse.

12—How to get the most out of memory editing

How to get the most out of memory editing

  • If you're brand new: Tell your companion the few things that matter most — what to call you, your schedule — then open /journal after a few messages and pin those facts.
  • If there's something she should never bring up: Add it on the Boundaries page, as a hard limit if you mean it absolutely.
  • If you've been chatting a while: Do one full read-through of each companion's page, and check About me too, since every companion can use those facts.
  • If you want depth: Edit preference and relationship facts, not just biographical ones. Adjusting how your companion treats a topic changes far more than correcting a date.
  • If a whole chapter is over: Forget the facts tied to it and keep the ones about you.

13—Final thoughts

Final thoughts

Editable memory is a small feature with a disproportionate effect on how a long-running AI relationship feels. It turns memory from something the system does to you into something you maintain together — and that shift is what lets a companion survive months of conversation instead of weeks.

It is also the least glamorous thing we've built. Nobody signs up for a text-editing interface. They stay because six months in, their companion still knows who they are.

FAQ

Can I see everything my AI companion remembers about me?

Every fact she has stored is in the Memory Journal as a plain sentence you can read. She can also draw on short snippets of your earlier messages, which come from your chat history rather than the journal.

Can I add a memory myself?

Boundaries, yes — on the Boundaries page. Other facts are captured from conversation, so say it clearly in chat, then edit or pin it once it appears in the journal.

Does editing a memory change past conversations?

No. Past messages stay as they were sent. Editing changes the fact used in future replies, so the effect shows up going forward rather than retroactively.

Is more memory always better?

Not really. Only a limited number of relevant facts go into each reply, so a well-curated journal, with the important facts pinned, often produces a sharper companion than a large, cluttered one.

Do I need to pay to use the Memory Journal?

No. The free tier gives you 20 messages a day, Aria, and the ability to create your own companions, and the Memory Journal is included. Premium adds unlimited messages, within fair use, Luna, Zara and Kai as well as Aria, all eight tones, group chat with two companions at once, and story worlds.

How is this different from other AI companion apps?

Our view is that a memory store you can read and correct is worth the trade-offs it introduces — you can see what your companion is working from, and fix it when it's wrong, instead of only seeing the effects.

15—A note from the team

A note from the team

This post is written by the team behind Svila.io. The features and choices we describe are ones we designed and shipped — so our perspective is first-party, not neutral. We try to be honest about the trade-offs, but you should always try things yourself and form your own view.


Last updated September 2026.

Ready to try Svila.io?

Create your first AI persona for free — no credit card required.

Get Started Free

Keep reading