Update to Kemory
Snapshot Oct 24, 2026 · 13:18 UTC · version 1.1.2
Package or technical metadata updated
Package or technical metadata updated
Observed in package metadata. These changes alone do not establish a new customer-facing feature.
Up to 3 plugins free · no card required.
All fields and original values.
These are saved file metadata. Open a changed file to read its text edits.
Package file
Oct 2, 2026 · 00:28 UTC
9fd18ebc96f08504373342b3eec686ea171ad285665c49d37b58943693023130
Oct 24, 2026 · 13:18 UTC
412fd948a86e5259cfb9721bf7b5934d07726d97332af389b9dbcb0e4013480b
Package file
Oct 2, 2026 · 00:28 UTC
127
Oct 24, 2026 · 13:18 UTC
98
Package file
Oct 2, 2026 · 00:28 UTC
3a953046cfd346b3a1b043a6f4e473910e641872da00d14cdcb2c95ee15e4271
Oct 24, 2026 · 13:18 UTC
6244dc76a17fab7ba56a091e8b2caf7c09931d6e7da564aa33c17f0517eb0e22
Package file
Oct 2, 2026 · 00:28 UTC
5319
Oct 24, 2026 · 13:18 UTC
1593
Compare saved observations
Download comparison JSONChanged files
Choose a file to see its text changes. Binary or large files show metadata only.
.codex-plugin/plugin.json
− removed · + added
--- before +++ after @@ -1,26 +1,36 @@ { - "apps": "./.app.json", + "name": "app-6a85afc169ac819185a36cd38c9fe432", + "version": "1.1.2", "author": { "name": "SeKondBrain.ai" }, - "description": "Every AI you talk to has amnesia, and a wall around it. You re-explain the same context to each one, every session, and none of them share what they learn.\n\nKemory is a permissioned memory that lives across every AI you use. Tell one thing to one assistant and they all understand, with you holding the keys. Connect it to ChatGPT and your preferences, project facts and decisions are there from the first message \u2014 including the ones you explained last week, in a different client, on a different machine.\n\nThis is not another memory inside one assistant. ChatGPT remembers what you tell ChatGPT. Kemory is the layer underneath that travels with you: the same memory answers in ChatGPT, in Codex, in Claude, in Cursor, and in any other MCP client \u2014 with organisation-level scoping and an audit trail that a single assistant's memory does not provide.\n\nWhat you can do with it:\n\n- Recall decisions, preferences and context from earlier conversations, without repeating them\n- Keep project knowledge in one place and draw on it across ChatGPT, Codex, Claude, Cursor and any other MCP client\n- Carry a thread from one AI into another without losing where you were\n- Inspect everything Kemory holds about you, and delete any of it\n\nHow it is governed: every memory is scoped to a user and an organisation, and that scoping is enforced in the database rather than in a prompt. Colleagues cannot read each other's memories by default.\nMemories are shared only when you deliberately promote them to a shared namespace, and every event lands in a hash-chained, tamper-evident audit log. We do not train on your data.\n\nKemory reads only what you deliberately capture into your own account. It never reads ChatGPT's own memory, chat history, conversation summaries or your files.\n\nKemory reads and writes. Every memory tool is titled and annotated, so ChatGPT knows reading from writing from deleting before it runs anything.\n\nRequires a SeKondBrain account. Sign in over OAuth; there is nothing to install and no keys to paste.\n\nKemory by SeKondBrain \u2014 the memory layer of the SeKondBrain Cognitive OS. Beta version.", + "description": "Consent-first persistent memory for your Kemory vault.", + "skills": "./skills/", + "mcpServers": "./.mcp.json", "interface": { - "capabilities": [], + "displayName": "Kemory", + "shortDescription": "Your private AI memory", + "longDescription": "Kemory retrieves relevant, user-specific context from your private vault and saves durable information only after you explicitly approve it. Use it to recover project decisions, preferences, and captured knowledge across sessions. It cannot access conversations that were not captured in Kemory, and it never stores credentials, secrets, or personal information unless you explicitly request retention.", + "developerName": "SeKondBrain.ai", "category": "Productivity", - "defaultPrompt": [ - "Set yourself up as my memory across assistants: save my role, my current project, and how I like answers formatted.", - "What do you already know about my projects and preferences? Load it before answering so I don't have to re-explain.", - "I worked through a decision in a different assistant. Pull up what I concluded and carry on from there." + "capabilities": [ + "Retrieve personal context", + "Search stored memories", + "Save consented memories" ], - "developerName": "SeKondBrain.ai", - "displayName": "Kemory", - "longDescription": "Every AI you talk to has amnesia, and a wall around it. You re-explain the same context to each one, every session, and none of them share what they learn.\n\nKemory is a permissioned memory that lives across every AI you use. Tell one thing to one assistant and they all understand, with you holding the keys. Connect it to ChatGPT and your preferences, project facts and decisions are there from the first message \u2014 including the ones you explained last week, in a different client, on a different machine.\n\nThis is not another memory inside one assistant. ChatGPT remembers what you tell ChatGPT. Kemory is the layer underneath that travels with you: the same memory answers in ChatGPT, in Codex, in Claude, in Cursor, and in any other MCP client \u2014 with organisation-level scoping and an audit trail that a single assistant's memory does not provide.\n\nWhat you can do with it:\n\n- Recall decisions, preferences and context from earlier conversations, without repeating them\n- Keep project knowledge in one place and draw on it across ChatGPT, Codex, Claude, Cursor and any other MCP client\n- Carry a thread from one AI into another without losing where you were\n- Inspect everything Kemory holds about you, and delete any of it\n\nHow it is governed: every memory is scoped to a user and an organisation, and that scoping is enforced in the database rather than in a prompt. Colleagues cannot read each other's memories by default.\nMemories are shared only when you deliberately promote them to a shared namespace, and every event lands in a hash-chained, tamper-evident audit log. We do not train on your data.\n\nKemory reads only what you deliberately capture into your own account. It never reads ChatGPT's own memory, chat history, conversation summaries or your files.\n\nKemory reads and writes. Every memory tool is titled and annotated, so ChatGPT knows reading from writing from deleting before it runs anything.\n\nRequires a SeKondBrain account. Sign in over OAuth; there is nothing to install and no keys to paste.\n\nKemory by SeKondBrain \u2014 the memory layer of the SeKondBrain Cognitive OS. Beta version.", - "privacyPolicyURL": "https://docs.sekondbrain.ai/legal/privacy/", - "shortDescription": "One memory. Every AI.", + "websiteURL": "https://sekondbrain.ai", "supportURL": "https://kemory.sekondbrain.ai/feedback", + "privacyPolicyURL": "https://docs.sekondbrain.ai/legal/privacy/", "termsOfServiceURL": "https://docs.sekondbrain.ai/legal/fair-usage/", - "websiteURL": "https://sekondbrain.ai" + "defaultPrompt": [ + "Find my saved decisions about this project.", + "What preferences have I saved in Kemory?", + "Help me save this decision to my Kemory vault." + ], + "brandColor": "#6D5DD3", + "brandColorDark": "#A88AF8", + "composerIcon": "./assets/kemory-icon.png", + "logo": "./assets/kemory-logo.png" }, - "name": "app-6a85afc169ac819185a36cd38c9fe432", - "version": "1.0.0" -} \ No newline at end of file + "apps": "./.app.json" +}
Full technical diff · 9 changed fields
changed /files/.app.json/sha256
"9fd18ebc96f08504373342b3eec686ea171ad285665c49d37b58943693023130"
"412fd948a86e5259cfb9721bf7b5934d07726d97332af389b9dbcb0e4013480b"
changed /files/.app.json/size
127
98
changed /files/.codex-plugin~1plugin.json/sha256
"3a953046cfd346b3a1b043a6f4e473910e641872da00d14cdcb2c95ee15e4271"
"6244dc76a17fab7ba56a091e8b2caf7c09931d6e7da564aa33c17f0517eb0e22"
changed /files/.codex-plugin~1plugin.json/size
5319
1593
added /files/.mcp.json
Field was absent
{
"sha256": "08b9fb81f743e19497ba47a430b3b7c2f46a07af7eb3c0f5f625c2db33f687d7",
"size": 96
}added /files/assets~1kemory-icon.png
Field was absent
{
"sha256": "fcc11ecdd59c6909e23405641dc9f28b6c689943d5f673c4e51d00fbddc2c36d",
"size": 34166
}added /files/assets~1kemory-logo.png
Field was absent
{
"sha256": "fcc11ecdd59c6909e23405641dc9f28b6c689943d5f673c4e51d00fbddc2c36d",
"size": 34166
}added /files/skills~1kemory-memory~1SKILL.md
Field was absent
{
"sha256": "12d942dbee5385f398a26f4469fd02c53a0811f9f74a89dc830980d848d1d831",
"size": 3664
}added /files/skills~1kemory-memory~1agents~1openai.yaml
Field was absent
{
"sha256": "feda231c547c74b2ccc4a9c7bb46266ea89fa5d0ddc61c53877d99204707b53e",
"size": 444
}Full snapshot data
{
"files": {
".app.json": {
"sha256": "412fd948a86e5259cfb9721bf7b5934d07726d97332af389b9dbcb0e4013480b",
"size": 98
},
".codex-plugin/plugin.json": {
"sha256": "6244dc76a17fab7ba56a091e8b2caf7c09931d6e7da564aa33c17f0517eb0e22",
"size": 1593
},
".mcp.json": {
"sha256": "08b9fb81f743e19497ba47a430b3b7c2f46a07af7eb3c0f5f625c2db33f687d7",
"size": 96
},
"assets/kemory-icon.png": {
"sha256": "fcc11ecdd59c6909e23405641dc9f28b6c689943d5f673c4e51d00fbddc2c36d",
"size": 34166
},
"assets/kemory-logo.png": {
"sha256": "fcc11ecdd59c6909e23405641dc9f28b6c689943d5f673c4e51d00fbddc2c36d",
"size": 34166
},
"skills/kemory-memory/SKILL.md": {
"sha256": "12d942dbee5385f398a26f4469fd02c53a0811f9f74a89dc830980d848d1d831",
"size": 3664
},
"skills/kemory-memory/agents/openai.yaml": {
"sha256": "feda231c547c74b2ccc4a9c7bb46266ea89fa5d0ddc61c53877d99204707b53e",
"size": 444
}
}
}SHA-256 of public snapshot: 9e5592724e526fb513c6efbdec7161a3ddcfda2956c59b45f140f02a1a792b97