ItzAditya43

Zenith — AI skill for Claude Code

AI community

Self-hosted, local-first AI workspace on your own Ollama models — chat, a Claude-Code-class coding agent, deep research, long-term memory, image gen, and voice.

How to install Zenith

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ItzAditya43/Zenith and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Zenith does

Self-hosted, local-first AI workspace on your own Ollama models — chat, a Claude-Code-class coding agent, deep research, long-term memory, image gen, and voice. No cloud, no accounts, nothing leaves your machine.

Alternatives in AI

Hindsight — State-of-the-art long-term memory for AI agents by Vectorize 6.7k ★
Cumora — Where agent teams gather 3.2k ★
Nocturne Memory — A lightweight, rollbackable, and visual Long-Term Memory Server for MCP Agents 1.3k ★

README

Zenith

A self-hosted, multimodal AI workspace built on your own Ollama instance — chat, a Claude-Code-class coding agent, deep research, memory, image generation, and voice, all running on your own hardware. No cloud calls except to your own local Ollama (and the open web, only when you explicitly ask it to search or attach a link).

Run it as a Docker stack, install it as a **native desktop app** (Tauri + frozen backend), or reach it from your phone over your LAN/Tailscale — same local backend, same data.

Everything in this document reflects what's actually implemented and has been verified against a real running deployment — not a roadmap, not aspirational copy.

This README is written for developers. If you're pointing someone non-technical at this project, [`docs/index.html`](docs/index.html) is a short, plain-language page explaining what it is and why — including a "Who it's for" section (private-ChatGPT user, developer, writer/researcher, privacy-first, tinkerer) that maps each audience to the handful of features that actually matter to them, rather than the full feature dump below. Open it directly, or enable it for free via GitHub Pages (Settings → Pages → deploy from `/docs` on this branch). The in-app onboarding tour asks the same question on first run and tailors its "try this first" step to the answer.