{"_id":"fleet-memory-mcp","_rev":"3-c66cc779850752af892e57c2bb8d5873","name":"fleet-memory-mcp","dist-tags":{"latest":"0.1.0"},"versions":{"0.0.0":{"name":"fleet-memory-mcp","version":"0.0.0","keywords":["mcp","memory","rcll","fleet-memory","hindsight","ai-agents","multi-agent","pgvector","long-term-memory","self-hosted","model-context-protocol"],"license":"MIT","_id":"fleet-memory-mcp@0.0.0","maintainers":[{"name":"geratron","email":"holetronn@gmail.com"}],"homepage":"https://rcll.ai","bugs":{"url":"https://github.com/holetron-lab/fleet-memory/issues"},"bin":{"fleet-memory-mcp":"server.js"},"dist":{"shasum":"debeed47f8deb86450d72055b5b9ae3661078572","tarball":"https://registry.npmjs.org/fleet-memory-mcp/-/fleet-memory-mcp-0.0.0.tgz","fileCount":4,"integrity":"sha512-nXncuDblr8MqM5FNJxOpKdQpP1tzqSBh0nWYXH+YPJrnmIiabnkuQe5VczSEzWCw3ibDMgn+3DQz95EU7FxmdQ==","signatures":[{"sig":"MEYCIQDndNe+iaIi+Xb/Ho9yjB2cjmi+2ny8cDECKEQz3mj0cwIhAPi9uj3+DVxQS4e9t22axo9obPA542uJ/BkpMKpprByz","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":18878},"main":"server.js","type":"module","engines":{"node":">=18"},"mcpName":"ai.rcll/fleet-memory","scripts":{"start":"node server.js"},"_npmUser":{"name":"geratron","email":"holetronn@gmail.com"},"deprecated":"Placeholder only — published solely to enable npm trusted publishing. The first real release is 0.1.0, built and signed by CI. See https://rcll.ai","repository":{"url":"git+https://github.com/holetron-lab/fleet-memory.git","type":"git"},"_npmVersion":"10.8.2","description":"RCLL — self-hosted, hierarchical shared memory for a team of AI agents over MCP (rooms, L0-L3 depth, pgvector). Fork of vectorize-io/hindsight.","directories":{},"_nodeVersion":"20.20.2","dependencies":{"@modelcontextprotocol/sdk":"^1.12.1"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/fleet-memory-mcp_0.0.0_1787899743636_0.2951372876123486","host":"s3://npm-registry-packages-npm-production"}},"0.1.0":{"name":"fleet-memory-mcp","version":"0.1.0","mcpName":"ai.rcll/fleet-memory","description":"RCLL — self-hosted, hierarchical shared memory for a team of AI agents over MCP (rooms, L0-L3 depth, pgvector). Fork of vectorize-io/hindsight.","homepage":"https://rcll.ai","repository":{"type":"git","url":"git+https://github.com/Holetron-lab/fleet-memory.git"},"bugs":{"url":"https://github.com/Holetron-lab/fleet-memory/issues"},"type":"module","main":"server.js","bin":{"fleet-memory-mcp":"server.js"},"scripts":{"start":"node server.js"},"keywords":["mcp","memory","rcll","fleet-memory","hindsight","ai-agents","multi-agent","pgvector","long-term-memory","self-hosted","model-context-protocol"],"license":"MIT","dependencies":{"@modelcontextprotocol/sdk":"^1.12.1"},"engines":{"node":">=18"},"gitHead":"77ea122dc92f10bfdf067a2f5d98799adc4371b3","_id":"fleet-memory-mcp@0.1.0","_nodeVersion":"22.23.2","_npmVersion":"12.0.2","dist":{"integrity":"sha512-MCl+Jd+0Bt4VDdTqqdUk/laQida9uHr884bKE9hrxyu/4ypxQg8ZfzTEEXnH7ULU62lipyVYgsQzN7vn7Csecw==","shasum":"aff95eb7833e8dcfca88e77b0d52e61fc7c1ed68","tarball":"https://registry.npmjs.org/fleet-memory-mcp/-/fleet-memory-mcp-0.1.0.tgz","fileCount":4,"unpackedSize":18877,"attestations":{"url":"https://registry.npmjs.org/-/npm/v1/attestations/fleet-memory-mcp@0.1.0","provenance":{"predicateType":"https://slsa.dev/provenance/v1"}},"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEYCIQCDbr6U0gVE18+QitJ1XEgGLW6s7O+imm43K6y8njLTQgIhAN92dpI8tl0Ls7+iJIdsaHrj0TLpHjFvqW4qrDGaAupa"}]},"_npmUser":{"name":"GitHub Actions","email":"npm-oidc-no-reply@github.com","trustedPublisher":{"id":"github","oidcConfigId":"oidc:c83d28c1-a5fd-4b22-8d23-629b17a35983"}},"directories":{},"maintainers":[{"name":"geratron","email":"holetronn@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/fleet-memory-mcp_0.1.0_1787920495929_0.7262897366347489"},"_hasShrinkwrap":false}},"time":{"created":"2026-08-28T06:49:03.509Z","modified":"2026-08-28T12:34:56.359Z","0.0.0":"2026-08-28T06:49:03.772Z","0.1.0":"2026-08-28T12:34:56.051Z"},"bugs":{"url":"https://github.com/Holetron-lab/fleet-memory/issues"},"license":"MIT","homepage":"https://rcll.ai","keywords":["mcp","memory","rcll","fleet-memory","hindsight","ai-agents","multi-agent","pgvector","long-term-memory","self-hosted","model-context-protocol"],"repository":{"type":"git","url":"git+https://github.com/Holetron-lab/fleet-memory.git"},"description":"RCLL — self-hosted, hierarchical shared memory for a team of AI agents over MCP (rooms, L0-L3 depth, pgvector). Fork of vectorize-io/hindsight.","maintainers":[{"name":"geratron","email":"holetronn@gmail.com"}],"readme":"# RCLL MCP Server (`fleet-memory-mcp`)\n\nStandalone [MCP](https://modelcontextprotocol.io) server that exposes **RCLL** memory tools to any MCP-compatible client (Claude Code, OpenClaw, etc). RCLL is self-hosted, hierarchical **shared** memory for a *team* of AI agents — topic-scoped rooms, L0–L3 depth, pgvector under the hood. The published artifact is named `fleet-memory`; RCLL is the product name.\n\n> RCLL — team memory for agent fleets. Built on Hindsight (github.com/vectorize-io/hindsight, MIT).\n\nCanonical repository: <https://godcrm.ai/git/holetron-lab/fleet-memory> — self-hosted, clonable\nanonymously. <https://github.com/holetron-lab/fleet-memory> is a read-only mirror of it, and is\nwhere issues and stars go.\n\nRCLL is a fork of [`vectorize-io/hindsight`](https://github.com/vectorize-io/hindsight) (MIT). It keeps Hindsight's storage engine and adds rooms — topic scoping over one shared store, which is selectivity rather than isolation — plus a hierarchical depth model (L0–L3). The room/hall/layer taxonomy is prior art in the hierarchical-memory space; the implementation here is our own.\n\n## Tools\n\n| Tool | Description |\n|------|-------------|\n| `memory_retain` | Save memories with room/hall/layer classification |\n| `memory_recall` | Scoped semantic search with room filtering |\n| `memory_reflect` | Deep reasoning + synthesis over stored memories |\n| `memory_compress` | Create closet summaries from accumulated facts |\n| `memory_bridge` | Cross-bank tunnels between related memories |\n\n## Quick Start\n\n```bash\nnpx fleet-memory-mcp\n```\n\n`FLEET_URL` points at your own RCLL backend (default `http://127.0.0.1:5100`); this package is the\nMCP client half and does not start a store for you. Standing one up is [the quickstart on\nrcll.ai](https://rcll.ai/docs/quickstart/).\n\nFrom a clone instead:\n\n```bash\ncd mcp-server\nnpm install\nFLEET_URL=http://localhost:5100 node server.js\n```\n\n## Claude Code\n\nAdd to `~/.claude/mcp.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"rcll\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"fleet-memory-mcp\"],\n      \"env\": {\n        \"FLEET_URL\": \"http://localhost:5100\",\n        \"FLEET_BANK\": \"my-agent-bank\"\n      }\n    }\n  }\n}\n```\n\n## Environment Variables\n\n| Variable | Default | Description |\n|----------|---------|-------------|\n| `FLEET_URL` | `http://127.0.0.1:5100` | fleet-memory backend base URL |\n| `FLEET_BANK` | `fleet-main` | Default memory bank ID. Set it explicitly. |\n\n### Migrating from `hindsight-mempalace-mcp`\n\nThat package defaulted to bank `mempalace-main`. `fleet-memory-mcp` defaults to `fleet-main`,\nso an install that never set the variable would open a different, empty bank — which\nreads as \"the update erased my memory\". It does not: the old bank is still there.\n\nSet `FLEET_BANK=mempalace-main` to keep reading it, or move the contents into a new\nbank first. When `FLEET_BANK` and `MEMPALACE_BANK` are both unset, the server prints\nwhich bank it defaulted to on stderr rather than picking one silently.\n\n### Deprecated (still read, with a notice on stderr)\n\nInstalls created before the rebrand keep working — these are used only when the\n`FLEET_*` equivalent is unset, and they will be dropped in a future major.\n\n| Legacy variable | Replaced by |\n|-----------------|-------------|\n| `HINDSIGHT_URL` | `FLEET_URL` |\n| `MEMPALACE_BANK` | `FLEET_BANK` |\n\n## Memory Taxonomy\n\n**Rooms** (topics): auth, pipeline, schema, infrastructure, ui, api, deployment, monitoring, agent, general\n\n**Halls** (knowledge types — the `hall` field): fact, event, decision, preference, discovery, procedure, warning\n\n**Layers** (depth):\n- L0 — Surface / identity (always at hand)\n- L1 — Critical (recalled by default)\n- L2 — Session (default for new memories)\n- L3 — Deepest burrow / archive (deep search only, compressed into closets)\n","readmeFilename":"README.md"}