By nomadsgalaxy
Two-way live-memory bridge to the Synaptic dashboard. v0.5.0 adds: (1) sd_pin_memory MCP tool + /sd-pin slash command for user-curated ground-truth memories (protected from dedup/decay, top-priority Phase 0b); (2) Stop hook now reads assistant text from transcript_path and ships it as response_complete with real content (was placeholder pulses only); (3) opt-in per-turn recall injection (SD_RECALL_INJECT=1) prepends top-K bank matches to every UserPromptSubmit. Existing hooks unchanged: SessionStart/End, UserPromptSubmit, PreToolUse, PostToolUse, SubagentStop, Stop. Bundled MCP server exposes 19 tools (added sd_pin_memory).
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Mark a Synaptic memory as sensitive (excluded from external research / Oracle / web fetch)
Pin a high-confidence, user-curated memory into Synaptic (protected from dedup/decay, deep-encoded at top priority)
Semantic-search the Synaptic memory bank for context relevant to your query
Trigger a Tier 3 Oracle web-research query through Synaptic (cached, budget-tracked, sensitive memories filtered)
One-shot Synaptic health check — token spend + recent activity + bank size
Matches all tools
Hooks run on every tool call, not just specific ones
Admin access level
Server config contains admin-level keywords
An agent-driven indexed memory store with a 3D MRI/CAT-scan-style dashboard. Memories are neurons in anatomical brain regions, shared tags become synapses, and live agent activity fires the regions in real time. Your agent queries the store; Synaptic remembers, organizes, and surfaces what's relevant.
If you want the short read on what Synaptic is and isn't, start with
docs/WHAT_SYNAPTIC_IS.md.
By NomadsGalaxy
Hello, I'm Nomad, and welcome to Synaptic. For a long time, I've been fascinated by how the brain works, how we think, store memories, conciousness as a whole. As I started to use AI more as a tool in my workflows, I wanted to better understand and visualize how it remembers things, I wanted to see what it would look like if I put my agent's brains in an MRI machine, and watch as the synapses fired, like lightning bolts in clouds.
So, I decided to try my hand at making it, admittedly heavily using Claude to generate it, and after several iterations, I want to publish it. What started as a visualizer ended up growing a full memory backend underneath — so it now does both: an indexed memory store your agent can write to and recall from, and the 3D brain dashboard that shows you what your agent is thinking. I hope you enjoy it as much as I do.
The best experience is to have the visualizer on a secondary monitor while using the plugin method like in Claude CLI, there are some configurations, so have fun with it.
The dashboard itself is dependency-free in the browser. Everything else is layered, so you only install the pieces you actually use. Docker is the convenience path, not a hard dependency — every component has a native install option, and the visualization runs end-to-end on the built-in mock engine with nothing more than Python + a browser.
| Component | Native requirement | Docker alternative | Required? |
|---|---|---|---|
SD Core + dashboard (single Go binary; SD Core serves index.html + assets at port 9911) | Go ≥ 1.25 (go build produces a static binary, no runtime deps) and a modern browser (Chrome / Edge / Firefox / Safari, recent versions) | synaptic-core container | Required unless you only want the mock dashboard at :8765 (see below) |
Mock-only dashboard (python serve.py, no SD Core) | Python 3.7+ and a modern browser | — (skip Docker entirely) | Optional, lower-fidelity. Brain runs on simulated activity; can't accept real adapter events. |
| Ollama (region classifier Tier 3 + thought bubbles) | Native install from https://ollama.com, then ollama pull llama3.2:3b | synaptic-ollama-tier1 container | Optional — falls back to keyword-only classification + pre-baked bubble pool |
| MCP adapter | Node.js ≥ 18 (run npm install in bridge/mcp-adapter/ once) | — (always native) | Only if you wire an MCP-aware client |
| OTel adapter | Node.js ≥ 18 (bridge/otel-adapter/) | — (always native) | Only if your runtime exports OpenTelemetry traces |
| Claude Code plugin | Nothing — Claude Code manages it via /plugin install | — (always in-process) | Only if you use Claude Code |
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub nomadsgalaxy/synaptic --plugin synaptic-claude-codeHarness-native ECC plugin for engineering teams - 67 agents, 278 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Evidence-gated AI coding workflow: scan → analyze → plan → TDD → execute → fix → verify → review, powered by Codebase Memory MCP >= 0.9.0 with optional Serena LSP intelligence. Includes blast-radius planning, test/cycle gates, independent review, and Windows Git Bash hook auto-resolution.
v9.52.0 - Reliability wave: tangle contextual review correction loop with hard round ceiling, progress-supervised review rounds (per-agent stall watch, descendant-tree kills), council diversity and agy pin fixes, marketplace generator source-of-truth fix, provider troubleshooting runbook and cost-expectations docs. Run /octo:setup.
Memory compression system for Claude Code - persist context across sessions
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.