superlocalmemory

Files (200)

• marketplace.json · .claude-plugin/
• marketplace.json · .cursor-plugin/
• .gitattributes
• FUNDING.yml · .github/
• marketplace.json · .github/plugin/
• installer-matrix.yml · .github/workflows/
• plugin-validate.yml · .github/workflows/
• pypi-publish.yml · .github/workflows/
• test.yml · .github/workflows/
• .gitignore
• .importlinter
• .lfsconfig
• .npmignore
• AGENTS.md
• ATTRIBUTION.md
• AUTHORS.md
• CHANGELOG.md
• CITATION.cff
• COMMERCIAL-LICENSE.md
• CONTRIBUTING.md
• LICENSE
• Makefile
• NOTICE
• PRIVACY.md
• README.md
• SECURITY.md
• SETUP_FOR_AGENTS.md
• SKILL.md
• _GENERATED.md · antigravity-plugin/
• slm-governance-advisor.md · antigravity-plugin/agents/
• slm-loop-runner.md · antigravity-plugin/agents/
• slm-memory-advisor.md · antigravity-plugin/agents/
• slm-optimize-advisor.md · antigravity-plugin/agents/
• slm-loop.md · antigravity-plugin/commands/
• hooks.json · antigravity-plugin/
• hooks.json · antigravity-plugin/hooks/
• mcp_config.json · antigravity-plugin/
• plugin.json · antigravity-plugin/
• antigravity_hook_adapter.py · antigravity-plugin/scripts/
• .gitattributes · assets/
• CONTACT-SHEET-README.md · assets/
• README.md · assets/
• README.md · assets/branding/
• google-g.svg · assets/branding/
• jev-mark.svg · assets/branding/
• slm-mark.png · assets/branding/
• slm-mark.svg · assets/branding/
• slm-wordmark-dark.svg · assets/branding/
• slm-wordmark-light.svg · assets/branding/
• contact-sheet.html · assets/
• contact-sheet.png · assets/
• favicon-white.ico · assets/
• .gitkeep · assets/gifs/
• QUICK-REFERENCE.md · assets/gifs/
• README.md · assets/gifs/
• USAGE.md · assets/gifs/
• cli-demo.gif · assets/gifs/
• dashboard-search.gif · assets/gifs/
• dashboard-tabs.gif · assets/gifs/
• event-stream.gif · assets/gifs/
• …and 140 more

SKILL.md (first lines)

# SuperLocalMemory

AI agent memory with a local data root. Five candidate producers (semantic, BM25, temporal, spreading-activation, Hopfield) fuse via RRF, with an entity-graph post-fusion score enhancement — all with mathematical similarity scoring. Mode A operates without sending memory content to a cloud model provider; optional connectors, backup, and proxy providers are explicit choices with separate behavior.

## Installation

```bash
pip install superlocalmemory
# or
npm install -g superlocalmemory
```

## Quick Start

```bash
slm remember "Alice works at Google as a Staff Engineer" --json
slm recall "Who is Alice?" --json
slm status --json
```

## Commands

All data-returning commands support `--json` for structured agent-native output.

### Memory Operations

```bash
slm remember "<content>" --json           # Store a memory
slm remember "<content>" --tags "a,b" --json
slm recall "<query>" --json               # Semantic search
slm recall "<query>" --limit 5 --json
slm list --json -n 20                     # List recent memories
slm forget "<query>" --json               # Preview matches (add --yes to delete)
slm forget "<query>" --json --yes         # Delete matching memories
slm delete <fact_id> --json --yes         # Delete specific memory by ID
slm update <fact_id> "<content>" --json   # Update a memory
```

### Pictures, documents and folders (4.1.25, optional)

Off by default. Needs a computer with 16 GB of memory and downloads about 1.5 GB.
Ask the user before turning it on.

```bash
slm media status --json                   # is it on, and how is set-up going
slm media enable                          # asks first; --yes skips the question (only with the user's yes)
slm media disable                         # memories are kept
slm media repair --dry-run                # count pictures that cannot be found by what they show yet
slm sources add ~/Notes --kind obsidian   # or --kind folder; read-only; shows what would be read, then asks
slm sources list --json
```

When it is on, a normal `slm recall` also returns saved pictures and PDF pages (a
`media` block per result). The MCP tools are `remember_media`, `remember_document`,
`get_media`, `media_status` and, for web apps, `media_upload_link`; they are listed
only in the `full`, `power` and `whole` tool sets. A picture is up to 25 MB, a PDF up
to 100 MB and 500 pages. Do not save a file that shows a secret. The `slm-media` skill
has the details.

### Company mode and the dashboard key

```bash
slm team status --json                    # Require login: on|off, user count
slm team policy --require-login on|off    # on the SLM computer; only when the user asks
slm token show                            # the key the dashboard asks for in strict mode; a secret, never paste it into chat
```

### Diagnostics

```bash
slm status --json                         # System status (mode, profile, DB)
slm health --json                         # Math layer health
slm trace "<query>" --json                # Recall with per-channel breakdown
```

### Configuration

```bash
slm mode --json                           # Get current mode
slm mode a --json                         # Set mode (a=local, b=ollama, c=cloud)
slm profile list --json                   # List profiles
slm profile switch <name> --json          # Switch profile
slm profile create <name> --json          # Create profile
slm connect --json                        # Auto-configure IDEs
slm connect --list --json                 # List supported IDEs
```

### Bounded Loops

```bash
slm loop demo                             # Run built-in convergence demo (no API key needed)
slm loop history [--name <loop-name>]     # List recorded runs from SLM memory
slm loop show <run_id>                    # Show every lap of one run
```

Loop laps are persisted to SLM memory under the tag `loop:<name>`. MCP tools
`slm_loop_run`, `slm_loop_history`, and `slm_loop_show` are available in the
`code` and `full` profiles.

### Services (no --json)

```bash
slm setup                                 # Interactive setup wizard
slm mcp                                   # Start MCP server (for IDE integration)
slm dashboard                             # Open web dashboard
slm warmup                                # Pre-download embedding model
```

## JSON Envelope

Every `--json` response follows a consistent envelope:

```json
{
  "success": true,
  "command": "recall",
  "version": "4.0.0",
  "data": {
    "results": [

From qualixar/superlocalmemory, AGPL-3.0.

More from qualixar/superlocalmemory