[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-79081":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":12,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":18,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":30,"readmeContent":31,"aiSummary":32,"trendingCount":14,"starSnapshotCount":14,"syncStatus":33,"lastSyncTime":34,"discoverSource":35},79081,"engram","NanoFlow-io\u002Fengram","NanoFlow-io","๐Ÿง  Hybrid long-term memory plugin for OpenClaw agents โ€” SQLite+FTS5 for structured facts, LanceDB for semantic recall",null,"TypeScript",120,1,119,0,8,0.9,"MIT License",false,"main",true,[22,23,24,25,26,27,28,29],"ai-agents","embeddings","lancedb","memory","openclaw","rag","sqlite","typescript","2026-06-12 02:03:49","# Engram\n\n```\n โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•—โ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ•—\n โ–ˆโ–ˆโ•”โ•โ•โ•โ•โ•โ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•”โ•โ•โ•โ•โ• โ–ˆโ–ˆโ•”โ•โ•โ–ˆโ–ˆโ•—โ–ˆโ–ˆโ•”โ•โ•โ–ˆโ–ˆโ•—โ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ–ˆโ–ˆโ•‘\n โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•— โ–ˆโ–ˆโ•”โ–ˆโ–ˆโ•— โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•‘ โ–ˆโ–ˆโ–ˆโ•—โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•”โ•โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•”โ–ˆโ–ˆโ–ˆโ–ˆโ•”โ–ˆโ–ˆโ•‘\n โ–ˆโ–ˆโ•”โ•โ•โ• โ–ˆโ–ˆโ•‘โ•šโ–ˆโ–ˆโ•—โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•‘ โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•”โ•โ•โ–ˆโ–ˆโ•—โ–ˆโ–ˆโ•”โ•โ•โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•‘โ•šโ–ˆโ–ˆโ•”โ•โ–ˆโ–ˆโ•‘\n โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•—โ–ˆโ–ˆโ•‘ โ•šโ–ˆโ–ˆโ–ˆโ–ˆโ•‘โ•šโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ•”โ•โ–ˆโ–ˆโ•‘ โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•‘ โ–ˆโ–ˆโ•‘โ–ˆโ–ˆโ•‘ โ•šโ•โ• โ–ˆโ–ˆโ•‘\n โ•šโ•โ•โ•โ•โ•โ•โ•โ•šโ•โ• โ•šโ•โ•โ•โ• โ•šโ•โ•โ•โ•โ•โ• โ•šโ•โ• โ•šโ•โ•โ•šโ•โ• โ•šโ•โ•โ•šโ•โ• โ•šโ•โ•\n\n hybrid long-term memory for OpenClaw agents\n```\n\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-yellow.svg)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT)\n[![OpenClaw](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FOpenClaw-plugin-purple)](https:\u002F\u002Fopenclaw.ai)\n\n> **engram** *(n.)* โ€” the physical\u002Fbiochemical trace a memory leaves behind. The thing that makes \"remembering\" actually possible.\n\nEngram is a memory plugin for [OpenClaw](https:\u002F\u002Fopenclaw.ai) agents. It gives them a brain that doesn't forget between sessions โ€” backed by **SQLite + FTS5** for exact, structured recall and **LanceDB** for fuzzy semantic search over embeddings.\n\n## โœจ Features\n\n- ๐Ÿง  **Hybrid recall** โ€” structured key\u002Fvalue facts + semantic vector search, queried together\n- ๐Ÿ” **FTS5 full-text search** over fact text\n- ๐Ÿ“š **Categories** โ€” `preference`, `fact`, `decision`, `entity`, `other`\n- โณ **Decay classes** โ€” `permanent`, `stable`, `active`, `session`, `checkpoint` (with confidence decay)\n- ๐Ÿช **Auto-capture \u002F auto-recall** hooks (configurable)\n- ๐Ÿ’พ **Local-first** โ€” your memory stays on your machine\n- ๐Ÿ”’ **Embeddings via OpenAI** (`text-embedding-3-small` or `-large`)\n\n## ๐Ÿš€ Install\n\nOne-liner (clones from this repo, copies the plugin into OpenClaw, wires the config):\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002FNanoFlow-io\u002Fengram\u002Fmain\u002Fscripts\u002Finstall.sh | bash\n```\n\nOr clone and run manually:\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FNanoFlow-io\u002Fengram.git\ncd engram\nbash scripts\u002Finstall.sh\n```\n\nAfter install:\n\n```bash\nexport OPENAI_API_KEY=\"sk-proj-...\" # if not already set\nopenclaw gateway restart\n```\n\nVerify it loaded โ€” look for `engram: initialized` in the gateway log.\n\n## ๐Ÿงช Try It\n\nFrom inside an OpenClaw agent session:\n\n```\n> Remember that I prefer dark mode and my coffee black.\n```\n\nNext session:\n\n```\n> What do you remember about me?\n```\n\nThe agent recalls via the `memory_recall` \u002F `memory_store` tools, both backed by Engram.\n\n## โš™๏ธ Configuration\n\nEngram is configured under `plugins.entries.engram` in `~\u002F.openclaw\u002Fopenclaw.json`:\n\n```json\n{\n \"plugins\": {\n \"slots\": { \"memory\": \"engram\" },\n \"entries\": {\n \"engram\": {\n \"enabled\": true,\n \"hooks\": { \"allowConversationAccess\": true },\n \"config\": {\n \"embedding\": {\n \"apiKey\": \"${OPENAI_API_KEY}\",\n \"model\": \"text-embedding-3-small\"\n },\n \"autoCapture\": true,\n \"autoRecall\": false,\n \"sqlitePath\": \"~\u002F.openclaw\u002Fmemory\u002Ffacts.db\",\n \"lanceDbPath\": \"~\u002F.openclaw\u002Fmemory\u002Flancedb\"\n }\n }\n }\n }\n}\n```\n\n| Field | Default | Notes |\n|---|---|---|\n| `embedding.apiKey` | โ€” | required (use `${OPENAI_API_KEY}`) |\n| `embedding.model` | `text-embedding-3-small` | or `text-embedding-3-large` |\n| `autoCapture` | `true` | extract memories from conversations |\n| `autoRecall` | `false` | inject relevant memories into context. Off by default โ€” can be noisy. |\n| `sqlitePath` | `~\u002F.openclaw\u002Fmemory\u002Ffacts.db` | structured store |\n| `lanceDbPath` | `~\u002F.openclaw\u002Fmemory\u002Flancedb` | vector store |\n\n## ๐Ÿ›  CLI\n\nEngram registers a `engram` CLI command with subcommands:\n\n```bash\nopenclaw engram stats # show fact\u002Fvector counts\nopenclaw engram prune # remove expired memories\nopenclaw engram checkpoint # manual checkpoint\nopenclaw engram backfill-decay # backfill decay classes on legacy data\nopenclaw engram extract-daily # extract from daily notes\nopenclaw engram search \u003Cq> # hybrid search\nopenclaw engram lookup \u003Ckey> # exact lookup\n```\n\nVerify after install:\n\n```bash\nopenclaw plugins doctor\nopenclaw plugins inspect engram --runtime\n```\n\n## ๐Ÿงฑ Architecture\n\n```\n โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”\n โ”‚ OpenClaw Agent โ”‚\n โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜\n โ”‚\n โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”\n โ”‚ Engram Plugin โ”‚\n โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”˜\n โ”‚ โ”‚\n โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ” โ”Œโ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”\n โ”‚ SQLite + FTS5 โ”‚ โ”‚ LanceDB โ”‚\n โ”‚ structured โ”‚ โ”‚ vectors โ”‚\n โ”‚ facts, keys, โ”‚ โ”‚ semantic โ”‚\n โ”‚ categories โ”‚ โ”‚ recall โ”‚\n โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜\n โ–ฒ โ–ฒ\n โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜\n โ”‚\n โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”\n โ”‚ OpenAI โ”‚\n โ”‚ embeddings โ”‚\n โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜\n```\n\nEvery memory gets written to **both** stores. Recall queries hit both and merge results by relevance.\n\n## ๐Ÿ—‘ Uninstall\n\n```bash\nbash scripts\u002Funinstall.sh\n```\n\nAsks before wiping memory data.\n\n## ๐Ÿค Contributing\n\nPRs welcome. Open an issue if you want to discuss architecture changes first.\n\n## ๐Ÿ“œ License\n\nMIT ยฉ NanoFlow\n\n---\n\n```\n โ–‘โ–‘โ–‘ remember everything. forget nothing. โ–‘โ–‘โ–‘\n```\n","Engram ๆ˜ฏไธ€ไธชไธบ OpenClaw ไปฃ็†่ฎพ่ฎก็š„ๆททๅˆ้•ฟๆœŸ่ฎฐๅฟ†ๆ’ไปถ๏ผŒๅฎƒ็ป“ๅˆไบ† SQLite + FTS5 ็”จไบŽ็ป“ๆž„ๅŒ–ไบ‹ๅฎžๅญ˜ๅ‚จๅ’Œ LanceDB ็”จไบŽ่ฏญไน‰ๅฌๅ›žใ€‚ๅ…ถๆ ธๅฟƒๅŠŸ่ƒฝๅŒ…ๆ‹ฌ็ป“ๆž„ๅŒ–้”ฎๅ€ผไบ‹ๅฎžไธŽ่ฏญไน‰ๅ‘้‡ๆœ็ดข็›ธ็ป“ๅˆ็š„ๆททๅˆๅ›žๅฟ†ใ€FTS5 ๅ…จๆ–‡ๆœ็ดขไปฅๅŠๅคš็ง่ฎฐๅฟ†็ฑปๅˆซๅ’Œ่กฐๅ‡็ฑปๅˆซ็š„ๆ”ฏๆŒใ€‚ๆญคๅค–๏ผŒ่ฏฅๆ’ไปถ่ฟ˜ๆไพ›ไบ†่‡ชๅŠจๆ•่Žท\u002F่‡ชๅŠจๅ›žๅฟ†้’ฉๅญ๏ผŒๅนถไธ”ไผ˜ๅ…ˆ่€ƒ่™‘ๆœฌๅœฐๅญ˜ๅ‚จไปฅ็กฎไฟๆ•ฐๆฎ้š็งใ€‚Engram ้€‚็”จไบŽ้œ€่ฆๅขžๅผบไปฃ็†่ฎฐๅฟ†่ƒฝๅŠ›็š„ๅบ”็”จๅœบๆ™ฏ๏ผŒๅฆ‚ไธชๆ€งๅŒ–ๆœๅŠกใ€ๆŒ็ปญๅฏน่ฏ็ณป็ปŸ็ญ‰๏ผŒ้€š่ฟ‡ OpenAI ็š„ๅตŒๅ…ฅๆจกๅž‹่ฟ›ไธ€ๆญฅๅขžๅผบไบ†่ฏญไน‰็†่งฃ่ƒฝๅŠ›ใ€‚",2,"2026-06-11 03:57:25","CREATED_QUERY"]