Claudian Plus

简体中文 | English

Your notes remember. Now your AI does too.
Six coding agents in one Obsidian vault, with a memory file you own.

Claudian Plus puts provider-backed coding agents in your Obsidian sidebar. Codex is the default; Claude, Kimi, OpenCode, Pi, and Antigravity plug into the same conversation model. When a session ends, its decisions do not evaporate — they are distilled into .claudian-plus/memory.md, a plain Markdown file inside your vault that you can read, edit, or delete.

01 / Watch it work

A 30-second animated overview of how the pieces fit together:

Click for the full-quality 30s 720p version.

Three things worth noticing

• Context arrives as references, not pasted text. @note, @folder, and drag-and-drop resolve against the vault you already have open.
• Memory lands in a file you can open. No database, no sync service — just memory.md sitting in your file list next to your notes.
• The provider switches without the conversation breaking. Each provider keeps its own session history, so a session can move between Codex, Claude, Kimi, and OpenCode.

02 / Why it is different

Most AI plugins treat conversation history as disposable text. Claudian Plus treats a session as a working record: context enters, decisions distill into memory, and the result stays searchable weeks later.

03 / What this is not

• Not a cloud service. No telemetry or analytics code ships in this plugin. Network calls are made by the provider CLI you configure, from your machine.
• Not a chat window bolted onto a note app. Your memory is a Markdown file in your vault, not a row in someone's database.
• Not a promise that every provider behaves the same. Capabilities differ per provider and are declared explicitly, because the codebase is written so no feature can silently assume parity.

04 / How memory actually works

Conversations end; memory continues. The pipeline is deliberately boring — four steps, no timers, and nothing written behind your back.

• It is gated on real usage, not on a stopwatch. Distillation is evaluated when a turn completes, and only once a session has at least 3 exchanges. A separate hourly check consolidates older logs when a full interval has passed, and startup consolidates anything accumulated while Obsidian was closed.
• Input is capped before the model sees it. The transcript handed to the auxiliary query is truncated at 8,000 characters, so distillation cost stays bounded on long sessions.
• Nothing is written silently. Newly extracted preferences and decisions surface as a notice and wait in Settings for you to accept.
• Output is capped before it reaches your prompt. Approved memory is injected within a 1,500 character budget by default, split across a 350-character profile layer and a 500-character project layer. All three are adjustable in Settings → Memory & Consciousness.
• You can always pull the plug. Open memory file shows you exactly what has been distilled, and the file is plain Markdown.

Memory is opt-in. memoryEnabled defaults to on, but the consciousness and auto-memory switches — the ones that send a transcript to an auxiliary query — default to off on new installations, because distilled memory contains personal context. Turn them on in Settings → Memory & Consciousness when you want it.

05 / Providers and the dependency rule

Three layers, one direction of dependency: vault context feeds a provider-neutral core, which dispatches to provider adaptors. Feature code depends on core contracts, never on provider internals.

• Codex CLI — enabled by default. Default model gpt-5.6-sol, with native streaming output over codex app-server.
• Claude Code — enabled by default. Thinking-chain folding, permission modes, and native project history replay.
• Kimi — opt in. ACP integration with dynamic model and command discovery, per-tool approval dialogs, and multimodal image attachments.
• OpenCode — opt in. ACP agent over a sandboxed sidecar process.
• Pi — opt in. RPC-mode sidecar that runs on Node built-ins and a bundled TypeBox, so it still works when external tool dependencies are missing.
• Antigravity — opt in, print mode only. Wraps the agy CLI with per-host CLI paths and a per-turn print timeout. Its capabilities are declared conservatively: only print mode is verified end-to-end, so persistent runtime, plan mode, rewind, fork, images, MCP, and shared skills are all reported as absent rather than assumed.

Kimi, OpenCode, Pi, and Antigravity are disabled on a fresh install and are enabled per provider in Settings. Only Codex and Claude start enabled.

06 / Install

Method 1 — from a release (recommended)

1. Download main.js, manifest.json, and styles.css from the latest release.
2. In your vault, create .obsidian/plugins/claudian-plus/.
3. Copy the three files into it.
4. In Obsidian, open Settings → Community plugins, refresh the list, and enable Claudian Plus.

Requirements: Obsidian 1.11.4 or later, desktop only. Claudian Plus talks to local agent CLI processes and the desktop filesystem, so there is no mobile build.

Method 2 — from source

Requires Node.js 24 and at least one provider CLI: Codex, Claude Code, Kimi, OpenCode, or Pi.

git clone https://github.com/wuyifan-code/Claudian-plus.git
cd Claudian-plus

npm ci
npm run typecheck
npm run build

Tip: set OBSIDIAN_VAULT=D:\Obsidian\My Vault in .env.local and npm run build will copy the build artifacts straight into your vault.

07 / Command reference

All sixteen commands registered by the plugin:

Conversation search lives in the sidebar itself rather than the command palette.

08 / Privacy and data layout

Everything the plugin knows lives inside your vault:

<your vault>/
├── .claudian-plus/
│   ├── claudian-plus-settings.json    # shared settings + per-provider config
│   ├── memory.md                      # distilled, injectable memory
│   └── sessions/
│       └── *.meta.json                # per-session metadata
├── .claudian/                         # legacy data, read non-destructively
└── .obsidian/plugins/claudian-plus/
    ├── main.js
    ├── manifest.json
    └── styles.css

• No telemetry. There is no analytics or tracking code in src/. Whatever leaves your machine leaves through the provider CLI or API endpoint you configured.
• No migration surprises. Legacy data under .claudian/ is detected and migrated without deleting the original.
• Settings writers merge. Provider-owned configuration is merged rather than overwritten, so switching versions does not wipe your setup.

09 / Development and verification

npm run dev                  # build CSS + esbuild watch
npm run typecheck            # TypeScript boundary checks
npm run lint                 # ESLint, including obsidianmd rules
npm run test                 # unit and integration suites
npm run test:watch           # re-run tests on change
npm run test:coverage        # coverage report
npm run test:architecture    # cross-layer dependency boundaries
npm run check:performance    # bundle ceiling + cold-evaluation timing

Two of these are worth knowing about before you open a pull request:

• test:architecture enforces the dependency rule in section 05 — feature code may not import provider internals.
• check:performance fails the build if main.js exceeds a 3.6 MB ceiling, and warns when median cold module evaluation exceeds a 50 ms indicator.

main.js is a build product and is not tracked. styles.css and versions.json are tracked for distribution — regenerate them with npm run build:css and npm run version rather than editing them by hand.

Start with AGENTS.md. It is the canonical cross-agent guide, and each scoped area under src/ has its own AGENTS.md with local rules.

10 / Built on

Claudian Plus is built on the work of two upstream projects:

Original and Claudian-derived code is licensed MIT; Codian-derived code retains AGPL-3.0 obligations. See LICENSE and NOTICE for file-level attribution.