Shared memory for Claude, Codex and your agents

One memory for every AI you use.

Give your AI apps and agents one long-term memory they all share. Tell one about a decision and the others already know it. It's plain Markdown you own, and setup is one URL.

Free tier. No card, no invite, no queue.

Aisle one / What it does

Remember. Share. Teach.

It isn't a chatbot or a notes app. It's one memory your AI apps connect to, so what you work out with one of them is there for all of them.

No. 1

It remembers

Your apps save what matters as you work: decisions, preferences, project facts, open questions. Next session they still know, so you stop re-explaining.

No. 2

Every app shares it

Claude, Codex and any app that speaks MCP read the same memory. Switch tools, or hit a usage cap, and your context comes with you.

NewNo. 3

It teaches your agents

What worked becomes a skill your agents load before they start. You approve every one, and each agent gets only the skills and permissions you give it.

Aisle two / Getting started

Three steps. The first one is a URL.

Nothing to install and nothing to import. You add one address to the AI apps you already use.

  1. Add one URL

    Paste it into your AI app's connector settings. Steps for Claude, Codex and others are below.

    https://openbrainstore.com/mcp
  2. Sign in with GitHub

    That's your account. The free tier starts right away: no card, no invite, no new password.

  3. Just work

    Your apps save what matters as you go and look it up when it helps. Ask one app what another learned.

Aisle three / See it work

Teach one. Ask another.

Two different apps, one account. A decision saved in one is an answer in the other.

In Claude · it saves

Remember: our retrieval uses vectors only as a fallback.

Saved as a decision, tagged retrieval and architecture.

In Codex · it recalls

What retrieval strategy did we choose?

Filters and full text first, then linked memories. Vectors only when those come up short.

Recalled from your memory, found by full text and links.

Same account. Different app. No copy and paste.

Aisle four / Skills New

Teach once. Every agent follows.

Memory is what your agents know. Skills are how you want things done: checklists, procedures and guardrails built from work that went well. Every skill is versioned, and none reaches an agent until you approve it.

  1. Step 1

    An agent proposes

    After a task goes well, an agent drafts a skill from it: your deploy checklist, how you review a pull request, a rule it should never break.

  2. Step 2

    You approve

    Drafts wait in the Catalog tab of the app. Edit, approve or reject them. Agents can never approve their own.

  3. Step 3

    Every agent loads it

    At the start of a session, each agent gets the skills that fit its task and fetches the ones it needs.

Each agent gets a role

Agent profiles give an agent its own role, skills and key, with only the permissions you pick. Change or revoke a key from the app; nothing new needs pasting into the agent.

Skills are opt-in: turn them on in the app's Catalog tab. Memory works the same either way.

Aisle five / Why it matters

Every app remembers a different you.

The more an AI app learns about you, the harder it is to leave. Your context turns into someone else's switching cost.

Today · a memory per app

Four apps, four memories. Switch tools and the story restarts: decisions disappear, and the briefing begins again.

With OpenBrainStore · one memory

Every app is a window onto the same memory. Your account decides what each one can reach, not the app.

Hit the cap. Keep the plot.

Max out one subscription, switch to another AI app, and keep going with the same project history. Your usage resets. Your context doesn't.

CLAUDEusage cap reached
CODEXsame project, picks up here
NEXT APPready when it ships

Aisle six / Finding things again

Just files? Exactly. That's the trick.

Any agent can drop notes in a folder. Finding them again is the hard part: ask about "auth decisions" and the note that says "OAuth and PKCE" never comes back.

OpenBrainStore keeps the same plain files underneath and does the librarian's job on top. Names and tags are tidied on the way in, so they agree no matter which agent wrote them. Every recall says which drawer found the answer.

Watch a recall think

Every recall opens these in order

Exact filters

Type, tags, people and projects, dates, key-values.

Full text

The words you actually wrote, matched fast.

Linked memories

Related notes one or two links away, like the reasoning behind a decision.

Deep search

Meaning, not just words, across everything you've stored. Opens on its own when the first three come up short, or whenever an agent asks.

Aisle seven / Yours to keep

Your memory. Your files.

Every memory is a plain Markdown file in an open format, the Open Knowledge Format, not a row locked inside one vendor's app. Read them, export them, or run the same server yourself.

retrieval-strategy.md
---
type: decision
tags: [retrieval, architecture]
entities: [openbrainstore]
links: [mem_01KXVZ…]
created: 2026-07-19
---
Retrieval runs structured filters and
full text first, then graph context.
Vectors run only as a fallback.
A memory, exactly as it's stored.
  • Plain Markdown

    Readable, diffable files with a few lines of front matter, indexed four ways so owning your files never means grepping them.

  • Works over MCP

    The open protocol AI apps already use to connect tools. No special SDK.

  • Tied to you

    Access belongs to your account, not to an app or a session, so every app reaches the same memory.

  • Portable by design

    Export the files whenever you like, or run the same server on infrastructure you control.

Aisle eight / Identity and trust

Your account holds the keys.

Every request is tied to your GitHub account. Apps and agents only get access you can see and take back.

Survives a rename

Your account is GitHub's numeric user ID, so changing your username never cuts you off from your memory.

Standard sign-in

Apps connect with OAuth 2.1 and PKCE. No client secrets to hand around.

Access you can revoke

Apps get rotating tokens. Agents get their own keys: limited to what you pick, set to expire, and changeable any time.

An exit that works

Export the Markdown or run the service yourself. The format doesn't depend on us.

Aisle nine / Two ways in

Own the box, or don't. Own the format.

Both paths share one open format and one open protocol, so there's no lock-in, even into us.

Self-hosted

Run your own instance, with full control and no cost beyond your own infrastructure. Free for individuals and companies under $1M a year.

Hosted, free to start

We run the same server for you, signed in with your GitHub account. The free tier starts the moment you sign in.

Aisle ten / Pricing

Two tags on the shelf.

No seats and no meters on recall. Export everything, on either plan, any time.

Free

$0forever

  • 1,000 memories
  • The full recall: filters, full text, links and deep search
  • Any MCP app, signed in with GitHub
  • Export any time
  • Open now: no invite, no card, no queue
Start free with GitHub
Opening soon

Pro

$6a month, or $60 a year

  • 50,000 memories: room for years of context
  • Everything in Free
  • Cancel any time. Downgrading never deletes your data
  • Keeps one indie project independent
Tell me when Pro opens

Aisle eleven / Connect an app

One URL. Then GitHub.

These are the real steps, and they work now. Signing in with GitHub creates your free account on the spot.

Claude

Open Settings, then Connectors, then Add custom connector. Leave Client ID and Secret empty. GitHub finishes the sign-in in your browser.

Name: openbrainstore
Remote MCP server URL:
https://openbrainstore.com/mcp

OAuth Client ID / Secret: leave empty

Aisle twelve / FAQ

Answers, without the fog.

Licensing, or just hello? Get in touch.

What is the Open Knowledge Format?

The open format underneath OpenBrainStore. Every memory is a Markdown file with a little front matter: typed, tagged, linked to people and projects, and tied to your account rather than to any AI vendor. You can read it, diff it, export it and run it anywhere. OpenBrainStore is the first OKF service: the memory layer.

How is this different from Obsidian or a Markdown vault?

Same files underneath, different job. A vault is for you to garden and search, and you carry context into your AI by hand. OpenBrainStore is a memory your AI apps write to mid-conversation and search for you, over MCP. You can still open every file; nobody has to garden them.

Is it just a vector database?

No. Exact filters and full text run first, linked memories add context, and deep semantic search runs when those come up short or an agent asks for it.

Does saving a memory use an LLM?

No. Tidying names, tags and types is deterministic. The memory itself is the plain text your app sends.

Can one app see what another saved?

Yes, when both are signed in as you. Your account decides what they can reach; the app and the session don't.

What are skills?

Reusable instructions for your agents: a checklist, a procedure, a guardrail. Agents propose them, you approve them in the app, and agents load the ones that fit their task at the start of a session. Assign skills to an agent profile and only that agent sees them. Skills are opt-in and versioned, and agents can never approve their own.

Do my agents need a key?

Apps like Claude and Codex sign in with GitHub and don't. Agents that run on their own, like scheduled jobs, CI or your own code, use a key from the app's Agents tab: limited to the permissions you choose, optionally bound to an agent profile, set to expire, and changeable or revocable any time.

What if I leave the hosted service?

Export the Markdown files and run the same server yourself. Local disk and S3-compatible storage keep the same portable structure.

What happens when I delete a memory?

It disappears from search and the graph right away. The file itself is kept for a short grace period before it's purged for good, so a mistake can be recovered.

Which apps work?

Claude and Codex are verified end to end. Other apps work if they support remote MCP servers with OAuth (Streamable HTTP).

Aisle thirteen / Get in touch

Licensing, or just hello.

OpenBrainStore is free to self-host for individuals and companies under $1M in yearly revenue (PolyForm Small Business license). Bigger, or want to offer it as a hosted service? Let's talk about a commercial license.

The next AI app is already coming.

Bring your memory with you when it does.