By apourmd941
Selran Hub — connect Claude to the local control room for the Selran ecosystem. Query your connected folders, databases, and APIs as MCP data sources, check connection health, and inspect schemas — all against the Hub daemon running privately on 127.0.0.1. Requires the free Selran Hub installed (pipx install git+https://github.com/apourmd941/selran-hub, or the signed macOS app). Local-first: no telemetry, no cloud backend.
The single local runtime for the Selran ecosystem — one daemon, one
well-known port (127.0.0.1:11999), one dashboard, one MCP bridge. By
Selran — written and maintained by Aidin
Eslampour.
Free Selran plugins (Greenloop, Design Director) work fully without it; the Hub adds optional enhanced experiences they detect and light up against (live clickable design pickers, the Greenloop audit panel, add-on packs).
Status: capabilities live — ports (registry), services (lifecycle with
verified process identity, incl. a safe orphan-PID reaper that finds and
stops leftover processes still holding an app's ports), studio (live
clickable pickers), audit (Greenloop live panel), entitlements (offline
pack licensing). The MCP bridge also exposes MCP Apps (SEP-1865) interactive
tools — a design-direction picker and a data-sources panel that render inline in
the chat on hosts that support them (with a text fallback elsewhere). As a Claude
Code plugin it also pushes events into your session: a SessionStart hook
injects the Hub's status as context, and a background monitor notifies you,
unprompted, when a local service goes down. Secrets (the API-key token,
data-source credentials) are stored in the OS keychain (macOS Keychain /
Windows Credential Manager / Linux Secret Service), with an owner-only 0600 file
fallback — never world-readable plaintext. selran-hub update-check tells you
when a newer release is out (opt-in automatic check; the only call it ever makes
out, and it sends nothing about you). The MCP bridge also exposes your connected
sources, schemas, and Hub health as read-only MCP resources (selran://…)
you can @-mention in Claude. For a Claude client on another machine, the bridge
can run a Streamable-HTTP transport — loopback by default; exposing it on a
network requires a bearer token (SELRAN_HUB_BRIDGE_TOKEN), enforced on every
request. Organizations can lock down a managed machine with a read-only
managed-policy.json (MDM-style) that the Hub enforces — restrict source types,
forbid remote access, require auth, disable update checks. Runs on macOS,
Windows, and Linux.
Local-first: binds
127.0.0.1only, no telemetry, no cloud backend. See PRIVACY.md. MIT licensed — see LICENSE.
The control room — data sources, ports, services, and packs in one place, with dark, light, and follow-the-system themes:
| Dark | Light |
|---|---|
![]() | ![]() |
Design Director packs get a gallery (each pack shows its real palette and type direction) and a multi-page example viewer — app screen, website, presentation, charts, and document, all rendered in the pack's own design language:
| Pack gallery | Example viewer |
|---|---|
![]() | ![]() |
macOS — signed installer (easiest). Download the latest
SelranHub-*.dmg from Releases,
drag to Applications, launch. Signed and notarized by Apple — no warnings. A
menu-bar icon shows status and opens the dashboard. (Windows and Linux
installers are in progress; use pipx below meanwhile.)
Any OS — pip / pipx (technical users):
pipx install git+https://github.com/apourmd941/selran-hub
selran-hub serve # run it (foreground)
selran-hub autostart install # or: start on login (LaunchAgent / systemd / Scheduled Task)
selran-hub status # check it
Claude Desktop — one-click bridge (.mcpb): build the MCP Bundle with
scripts/build_mcpb.sh and drag dist/selran-hub-*.mcpb into Claude Desktop →
Settings → Extensions. This installs the bridge that exposes your Hub's tools
to Claude; the Hub daemon still installs via one of the options above. See
docs/MCPB.md.
Developers (from a clone): ./bootstrap.sh then ./run_hub.sh. Tests:
python3 -m pytest tests/ -q (in-process, no bound ports).
See SELRAN_HUB_DESIGN.md for the full specification and roadmap, and ops/P3-PLAN.md for the installer build pipeline.
Bug reports and feature requests are welcome — open an issue. Pull requests are not accepted: to keep authorship and licensing unambiguous, all code here is written by the author. Describe a fix in an issue and it'll be credited.
MIT — © 2026 Selran, Aidin Eslampour. Privacy: PRIVACY.md. Contact: [email protected]
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 claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub apourmd941/selran-hub --plugin selran-hubSelran Greenloop — the loop that runs until green. Point it at any codebase and it maps the architecture, reconstructs the spec, audits against ten quality categories, fixes what it finds, and adversarially re-verifies — repeating until nothing is left to fix. Five skills: cartographer, spec-bootstrap, pre-commit-verification, app-audit, audit-fix. Free for personal, non-commercial use.
context-keeper — cross-session memory for Claude Code. Recall ranked, token-budgeted context from every past session (recall, list_sessions, list_projects MCP tools), with an optional auto-recall hook and a local mind-map UI of your projects, topics, and sessions. Local-first: your transcripts are indexed on your machine by a local embedding model; nothing is uploaded. Requires the free context-keeper daemon (cargo install or repo build).
MCP server that saves 98% of your context window with session continuity. Sandboxed code execution in 11 languages, FTS5 knowledge base with BM25 ranking, and automatic state restore across compactions.
Open-source, local-first Claude Code plugin for token reduction, context compression, and cost optimization using hybrid RAG retrieval (BM25 + vector search), reranking, AST-aware chunking, and compact context packets.
Give your AI a memory — mine projects and conversations into a searchable palace. 33 MCP tools, auto-save hooks, and guided setup.
Complete AI coding workflow system. Self-correcting memory + persistent FTS5-indexed research wikis + auto-research loop + multi-LLM council on a single SQLite store. 33 skills, 8 agents, 22 commands, 37 hook scripts across 24 events. Cross-agent via SkillKit.
Write SQL, explore datasets, and generate insights faster. Build visualizations and dashboards, and turn raw data into clear stories for stakeholders.
Agent Skills for AI/ML tasks including dataset creation, model training, evaluation, and research paper publishing on Hugging Face Hub