AttiClaw

Personal AI model management and inference platform, built on OpenClaw.

Version: 2026.2.21 | Docs: docs.openclaw.ai | Discord: discord.gg/qkhbAGHRBT | License: MIT

AttiClaw is your own standalone web dashboard for browsing, downloading, and running AI models locally. It pulls features from multiple upstream open-source projects (OpenClaw, ClawX, llama.cpp, IronClaw, ZeroClaw, and more) while maintaining a clean fork-based architecture.

Architecture Overview
Quick Start
Installation (Ubuntu)
AttiClaw Web App
Storage Architecture
Model Management
Upstream Feature Sources
OpenClaw Core
Council System
Memory & RAG
Smart Router
Inference Backends
Skills & ClawHub
Extensions
Voice & Media
Personality & Identity
Security
Database Backends
Deployment
Project Structure
Configuration
Testing
CLI Reference
Contributing
License

Architecture Overview

                        YOUR APP
                    ┌──────────────┐
                    │   AttiClaw   │  ← Standalone web dashboard
                    │  (React+Vite │     Port 5180
                    │   ShadCN/UI) │     Your own branding & features
                    └──────┬───────┘
                           │
              ┌────────────┼────────────┐
              │            │            │
              v            v            v
     ┌────────────┐ ┌───────────┐ ┌──────────┐
     │  OpenClaw   │ │  Storage  │ │  Models  │
     │  Gateway    │ │ SMB+NVMe  │ │ HuggingFace│
     │  ws://3100  │ │ Client    │ │ Hub Client │
     └──────┬─────┘ └─────┬─────┘ └─────┬────┘
            │              │              │
     ┌──────┴──────┐  ┌───┴───┐   ┌──────┴──────┐
     │ Agent Loop  │  │  SMB  │   │ Download →  │
     │ Council     │  │  (all │   │ SMB (store) │
     │ Memory/RAG  │  │ files)│   │ NVMe (run)  │
     │ Skills      │  └───────┘   └─────────────┘
     └──────┬──────┘
            │
     ┌──────┴──────────────────────────────────┐
     │           Inference Backends              │
     │  Ollama │ llama.cpp │ vLLM │ LM Studio   │
     └─────────────────────────────────────────┘

     UPSTREAM FEATURE SOURCES (git submodules in repos/):
     ┌─────────────────────────────────────────┐
     │ clawx      │ ironclaw  │ llama.cpp      │
     │ nanobot    │ picoclaw  │ zeroclaw       │
     │ awesome-openclaw-usecases               │
     └─────────────────────────────────────────┘

Key principles:

AttiClaw is YOUR app — independent from upstream projects
OpenClaw is the core engine (upstream fork)
All feature sources are git submodules with upstream remotes
SMB = permanent storage for everything; NVMe = fast cache for running models only
You choose which models to use; the system suggests better ones; you approve

Quick Start

# Clone
git clone https://github.com/attilaczudor/Test.git atticlaw
cd atticlaw

# Initialize all upstream submodules
./setup-repos.sh

# Install and build OpenClaw core
pnpm install && pnpm build

# Install and start AttiClaw
cd AttiClaw && pnpm install && pnpm dev
# Open http://localhost:5180

# In another terminal — start OpenClaw backend
cd .. && pnpm start

Installation (Ubuntu)

Full automated installer for Ubuntu 22.04 / 24.04:

sudo bash install-ubuntu.sh

The installer handles everything:

Step	What it does
1. System deps	build-essential, curl, git, python3, jq
2. Storage tools	cifs-utils (SMB), nvme-cli, mount points
3. Node.js 22	Via NodeSource
4. pnpm	Via corepack
5. Clone repo	Git clone + checkout
6. Submodules	All 8 upstream feature sources
7. OpenClaw core	Dependencies + TypeScript build
8. AttiClaw app	Dependencies + Vite build
9. Ollama	LLM inference backend

Plus: systemd services (openclaw.service, atticlaw.service), SMB mount guidance.

# Options
sudo bash install-ubuntu.sh --headless     # Unattended, all defaults
sudo bash install-ubuntu.sh --dir ~/ai     # Custom install path
sudo bash install-ubuntu.sh --skip-ollama  # Skip Ollama
sudo bash install-ubuntu.sh --skip-smb     # Skip SMB tools

Manual Install

# Prerequisites: Ubuntu, Node.js >= 22, pnpm
git clone https://github.com/attilaczudor/Test.git /opt/atticlaw
cd /opt/atticlaw

# Submodules
git submodule update --init --recursive
./setup-repos.sh

# OpenClaw core
pnpm install && pnpm build

# AttiClaw
cd AttiClaw && pnpm install && pnpm build

# LLM backend
curl -fsSL https://ollama.ai/install.sh | sh
ollama pull llama3.2:3b

AttiClaw Web App

AttiClaw is a standalone React web application — your personal AI dashboard.

Technology Stack

Technology	Purpose
React 19	UI framework
Vite	Build tool + dev server (port 5180)
TypeScript	Type safety
Tailwind CSS	Utility-first styling
ShadCN/UI	Component library (12 components)
Zustand	State management with localStorage persistence
react-i18next	Internationalization (EN, ZH, JA)
React Router	Client-side routing
Sonner	Toast notifications

Pages

Page	Description
Dashboard	System overview — models stored, models running, SMB status, NVMe usage
Chat	Conversation interface (connects to OpenClaw agent)
Models	HuggingFace browse, download, recommendations, approve workflow, storage config
Settings	Theme, language, SMB config, NVMe drive selector

ShadCN/UI Components

Button, Card, Badge, Tabs, Input, Select, Dialog, Progress, Separator, Switch, Tooltip, Label — all following the ShadCN convention with Radix UI primitives.

Development

cd AttiClaw
pnpm install
pnpm dev          # Dev server on http://localhost:5180
pnpm build        # Production build to dist/
pnpm typecheck    # TypeScript checking

Storage Architecture

Two-tier storage with clear separation of concerns:

┌─────────────────────────────────────────────┐
│              SMB Share (permanent)            │
│  //192.168.1.100/models                      │
│                                              │
│  models/          ← all downloaded models    │
│  datasets/        ← training data, samples   │
│  voice-samples/   ← voice cloning refs       │
│  configs/         ← model configs            │
│  files/           ← other files              │
└──────────────────────┬──────────────────────┘
                       │ Load (copy)
                       v
┌─────────────────────────────────────────────┐
│              NVMe Drive (runtime cache)      │
│  /mnt/nvme-models                            │
│                                              │
│  ONLY running models live here               │
│  Auto-evicts LRU when full                   │
│  Deleted on unload (stays on SMB)            │
└─────────────────────────────────────────────┘

Storage	Purpose	Persistence	Access speed
SMB	All files — models, datasets, configs	Permanent	Network
NVMe	Running models only	Temporary cache	Fast local
Cache	Fallback when SMB disconnected	Until SMB connects	Local disk

Key behaviors

Downloads always target SMB (permanent storage)
Models copied to NVMe only when loaded for inference
Unload removes NVMe copy — model stays safe on SMB
Auto-evicts least-recently-used from NVMe when capacity is low
Auto-migrates cached files to SMB when the share reconnects
Non-model files (datasets, voice samples) never touch NVMe
SMB credentials configurable in AttiClaw Settings UI

SMB Configuration

Configure via AttiClaw Settings page or /etc/fstab:

# Credentials file
echo -e 'username=YOUR_USER\npassword=YOUR_PASS' > /root/.smb-credentials
chmod 600 /root/.smb-credentials

# /etc/fstab entry
//192.168.1.100/models  /mnt/smb-models  cifs  credentials=/root/.smb-credentials,uid=1000  0  0

Model Management

HuggingFace Hub Integration

Browse and download models from HuggingFace directly in AttiClaw:

Feature	Description
Search	Search HuggingFace by name, task, format
All formats	GGUF, SafeTensors, PyTorch, ONNX, GPTQ, AWQ
All tasks	Text generation, speech-to-speech, TTS, STT, vision, embeddings
Runtime mapping	Auto-detects which backends can serve each format
Sharded models	Detects and handles multi-file sharded models
Download progress	Real-time progress with speed and ETA

Format → Runtime Compatibility

Format	Compatible Runtimes
GGUF	Ollama, llama.cpp, LM Studio
SafeTensors	vLLM, transformers, TGI
PyTorch	transformers
ONNX	ONNX Runtime
GPTQ / AWQ	vLLM, transformers

Model Recommendations

The system suggests models based on your hardware and needs. You approve or dismiss each suggestion.

Category	Recommended Models
General Chat	Qwen3-32B-GGUF (95), Llama-3.1-8B-GGUF (90), Llama-3.3-70B (96)
Code	Qwen2.5-Coder-32B (95), Qwen2.5-Coder-7B-GGUF (88)
MoE	Kimi-K2.5 (94) — requires vLLM + multi-GPU
Voice	personaplex-7b (92, speech-to-speech), whisper-large-v3-turbo (93, STT), fish-speech-1.5 (87, TTS)
Multimodal	llama3-llava-next-8b (85, vision+text)
Embeddings	bge-m3 (94, multilingual), nomic-embed-text-v1.5 (88, 8K context)

Workflow

Browse HuggingFace → Select model → Download to SMB
                                          ↓
     System suggests better model → You approve/dismiss
                                          ↓
     Load to NVMe (fast cache) → Assign to backend → Inference
                                          ↓
     Stop model → Unload from NVMe → Stays safe on SMB

Upstream Feature Sources

All upstream repos are git submodules with fork + upstream remotes:

Submodule	Upstream	Description
`repos/clawx`	ValueCell-ai/ClawX	Electron desktop UI (React + Zustand + i18n)
`repos/ironclaw`	nearai/ironclaw	Near AI agent infrastructure, distributed orchestration
`repos/llama.cpp`	ggml-org/llama.cpp	GGUF model inference engine
`repos/zeroclaw`	zeroclaw-labs/zeroclaw	Zero-shot agent framework
`repos/nanobot`	HKUDS/nanobot	Lightweight AI agent (HKU Data Science)
`repos/picoclaw`	sipeed/picoclaw	Edge deployment on RISC-V hardware
`repos/artemis`	Stanford-Trinity/ARTEMIS	AI research framework (Stanford)
`repos/awesome-openclaw-usecases`	anthropics/awesome-openclaw-usecases	29 curated use cases (3.8k stars)

Managing Upstream Repos

# Initialize all submodules + set upstream remotes
./setup-repos.sh

# Fetch latest from all upstreams
./setup-repos.sh --update

# Show status
./setup-repos.sh --status

# Merge upstream changes into a specific repo
cd repos/ironclaw && git fetch upstream && git merge upstream/main

All repos use the fork pattern: attilaczudor/<name> (fork) ← <org>/<name> (upstream).

OpenClaw Core

The core engine (upstream fork of openclaw/openclaw):

Agentic reasoning — Multi-turn loop with tool use, memory, shell execution
Local-first — Ollama, vLLM, llama.cpp, LM Studio
Council-gated escalation — Cloud only when local models lack confidence + user approval
Graph memory — Neo4j-backed knowledge graph with importance scoring and decay
Hybrid RAG — BM25 + vector similarity + graph traversal
LoRA fine-tuning — Continuous learning from high-quality interactions
53 MCP skills — Productivity, dev, communication, media, AI/ML, system
37 extensions — Discord, Slack, Telegram, WhatsApp, Signal, Matrix, IRC, and more
Real-time voice — Full-duplex with VAD and barge-in
Multi-platform — Node.js, Electron, macOS, iOS, Android

Council System

3-tier hierarchical council for complex reasoning:

Tier	Role	Params	Count	Description
T1	Director	3-70B	1	Decomposes, synthesizes, manages escalation
T2	Branch	2-20B	2-5	Any expertise, dispatches to specialists
T3	Specialist	0.5-7B	Up to 10/branch	Small focused models, parallel answers

Every member has persistent memory (vector DB, graph, LoRA) across model swaps.

Deliberation: Prompt → Director decomposes → Branches dispatch → Specialists answer in parallel → Branches consolidate → Director synthesizes → Cloud escalation if needed (user approval required).

Memory & RAG

Graph Memory — Neo4j knowledge graph with node types (fact, task, contact, file, experience, summary), importance scoring (0-1), decay, auto-summarization
Vector Store — Qdrant embeddings with cosine similarity, configurable dimensions
Hybrid RAG — BM25 keyword + vector similarity + graph traversal, local embeddings

Smart Router

Local-first with intelligent escalation:

Query → Local LLM → Confidence >= 0.6? → Return
                         ↓ NO
         Council Deliberation → Cloud needed?
                                    ↓ YES
                    User Approval → Cloud API → Return

Cost controls: per-request limit ($0.50), daily budget ($10), per-provider tracking.

Inference Backends

Backend	URL	Formats	Description
Ollama	localhost:11434	GGUF	Full-featured, model pulling, GPU offload
llama.cpp	localhost:8080	GGUF	Direct GGUF serving
vLLM	localhost:8000	SafeTensors, GPTQ, AWQ	High-throughput, PagedAttention
LM Studio	localhost:1234	GGUF	GUI-based, OpenAI-compatible
transformers	—	All	Python-based, universal
TGI	—	SafeTensors	Text Generation Inference
ONNX Runtime	—	ONNX	Optimized inference

Autodiscovery probes all known ports on startup.

Skills & ClawHub

53 bundled MCP skills across productivity, dev, communication, media, AI/ML, system, hardware, web, and home categories.

ClawHub marketplace with SHA-256 verification, VirusTotal scanning (mandatory after ClawHavoc incident — 341 malicious skills removed Feb 2026).

Extensions

37 extensions: Discord, Slack, Telegram, WhatsApp, Signal, iMessage, Matrix, IRC, LINE, Nostr, Twitch, MS Teams, Google Chat, Feishu, Mattermost, Nextcloud Talk, Zalo, Tlon, voice-call, phone-control, copilot-proxy, diagnostics-otel, and more.

Voice & Media

Real-time voice — Full-duplex with VAD, barge-in, continuous loop
TTS — Piper (CPU) / NVIDIA NeMo Parakeet (GPU)
STT — Whisper.cpp (CPU) / NVIDIA NeMo Canary (GPU)
Vision — LLaVA via Ollama
Voice models — nvidia/personaplex-7b (speech-to-speech), fish-speech (TTS with cloning)

Personality & Identity

Persistent personality system (PersonalityDb): AI name/style/traits, user profile tracking, interaction pattern learning, conversation summarization. Multi-backend storage (SQL + Neo4j + Qdrant).

Security

Layer	Protection
Sandbox	Wasm/Docker/nsjail with memory+CPU limits
RBAC	17+ permission types, path-level restrictions
CSRF	Ed25519 signed tokens on WebSocket messages
Origins	Strict allow-list (CVE-2026-25253 mitigation)
Rate Limiting	Per-IP sliding window
TLS	Optional HTTPS/WSS
Skills	Signature verification + VirusTotal scanning
Downloads	SHA-256 hash verification
Hashing	SHA-256 for gateway lock and tool-call IDs (migrated from SHA-1)
Secrets	`detect-secrets` scanning in CI/CD
Docker	Base images pinned to SHA-256 digests, non-root runtime user
Node.js	Requires >= 22.12.0 (CVE-2025-59466, CVE-2026-21636 patched)

Recent hardening: config validation with integer overflow protection, shell injection prevention, type-safe error handling, and concurrency guards. See SECURITY.md for reporting and operational guidance.

Database Backends

Database	Purpose	Development	Production
Qdrant	Vector embeddings (RAG)	In-memory	VM 10.0.0.50:6333
Neo4j	Knowledge graph memory	In-memory	VM 10.0.0.51:7474
MariaDB	Relational (repos, KV, personality)	SQLite	VM 10.0.0.52:3306

Deployment

Proxmox Layout

Proxmox Host (Threadripper, 256 GB RAM)
│
├── LXC 301: openclaw (Gateway + Agent)
│   ├── ws://10.0.0.100:3100  (Gateway)
│   ├── http://10.0.0.100:3101 (Web UI)
│   └── http://10.0.0.100:5180 (AttiClaw)
│
├── VM  302: council-director (mixtral:8x7b, 32 GB)
├── LXC 303-309: council members (T2+T3)
│
├── VM  310: qdrant (vector DB, 4 GB)
├── VM  311: neo4j (graph DB, 8 GB)
└── VM  312: mariadb (SQL, 4 GB)

Docker Compose

docker compose up -d    # Full stack with Qdrant, Neo4j, MariaDB

Systemd

sudo systemctl start openclaw    # Gateway + Agent
sudo systemctl start atticlaw    # Web Dashboard

Project Structure

.                           Root (fork of openclaw/openclaw)
├── AttiClaw/               YOUR APP — standalone web dashboard
│   ├── src/
│   │   ├── pages/          Dashboard, Chat, Models, Settings
│   │   ├── stores/         Zustand stores (models, settings)
│   │   ├── components/     ShadCN/UI + layout (sidebar, etc.)
│   │   └── i18n/           EN, ZH, JA translations
│   ├── package.json        Independent from OpenClaw
│   └── vite.config.ts      Dev server on port 5180
│
├── repos/                  Upstream feature source submodules
│   ├── clawx/              ClawX desktop UI (Electron)
│   ├── ironclaw/           Near AI agent infrastructure
│   ├── llama.cpp/          GGUF inference engine
│   ├── zeroclaw/           Zero-shot agent framework
│   ├── nanobot/            Lightweight agent
│   ├── picoclaw/           Edge deployment (RISC-V)
│   ├── artemis/            ARTEMIS AI research (Stanford)
│   └── awesome-openclaw-usecases/
│
├── src/                    OpenClaw core (88+ modules)
│   ├── agent/              Reasoning loop
│   ├── agents/             Agent system (auth, sandbox, tools, skills)
│   ├── auto-reply/         Automated response handling
│   ├── channels/           Communication channel framework
│   ├── cli/                CLI interfaces (gateway, daemon, cron, nodes)
│   ├── commands/           Command handlers
│   ├── config/             Configuration management
│   ├── council/            3-tier LLM council
│   ├── gateway/            WebSocket server + protocol
│   ├── hooks/              Lifecycle hooks
│   ├── infra/              HuggingFace hub + storage client
│   ├── memory/             Graph memory + Neo4j
│   ├── pairing/            Device pairing
│   ├── personality/        Personality system
│   ├── plugins/            Plugin runtime
│   ├── providers/          Multi-provider LLM
│   ├── rag/                Hybrid RAG pipeline
│   ├── rbac/               Role-based access control
│   ├── router/             Smart Router (council-gated)
│   ├── sandbox/            Execution sandbox
│   ├── security/           Security subsystem
│   ├── sessions/           Session management
│   ├── skills/             MCP skill interface
│   ├── tui/                Text UI interface
│   ├── voice/              Real-time voice
│   └── ...                 (88+ modules total)
│
├── Swabble/                Swift SDK package
├── extensions/             37 channel/media extensions
├── skills/                 53 bundled MCP skills
├── apps/                   Native apps (macOS, iOS, Android)
├── packages/               Workspace packages (clawdbot, moltbot)
├── ui/                     Web UI (Vite + Vitest)
├── docs/                   Documentation site (docs.openclaw.ai)
├── deploy/                 Proxmox automation scripts
├── scripts/                Build and utility scripts
├── test/                   E2E and integration tests
├── assets/                 Icons, images, chrome extension
├── vendor/                 Vendored dependencies (a2ui)
├── git-hooks/              Pre-commit hooks
│
├── install-ubuntu.sh       Ubuntu full-stack installer
├── install.sh              Universal installer
├── setup-repos.sh          Submodule init + upstream remotes
├── openclaw.json           Runtime configuration (generated)
└── .gitmodules             Submodule declarations

Configuration

openclaw.json — generated by pnpm openclaw init:

{
  "version": "2.0.0",
  "useCloudModels": false,
  "agent": { "defaultModel": "llama3.2:3b", "maxTurns": 100 },
  "gateway": {
    "host": "127.0.0.1",
    "port": 3100,
    "allowedOrigins": ["http://localhost:3101", "http://localhost:5180"],
  },
  "memory": { "backend": "hybrid", "maxNodes": 10000 },
  "sandbox": { "enabled": true, "runtime": "wasm" },
  "discovery": { "enabled": true },
}

See src/config/schema.ts for full JSON Schema.

Testing

1,375+ test files across src/, test/, and extensions/:

pnpm test                # Full suite (parallel, Vitest)
pnpm test:fast           # Unit tests only
pnpm test:coverage       # With V8 coverage
pnpm test:all            # Full CI pipeline: lint, build, unit, e2e, live, docker

Multiple Vitest configurations for different scopes: unit, e2e, gateway, extensions, and live tests. Coverage thresholds enforced at 70% lines/functions/statements and 55% branches.

Covers: config, council, CSRF, rate limiting, graph memory, Neo4j, SQL, RAG, RBAC, sandbox, skills, vector store, smart router, ClawHub, HuggingFace hub, personality, providers, agents, gateway, Docker, and more.

Build Toolchain

Tool	Purpose
TypeScript ^5.9.3	Type checking
tsdown ^0.20.3	Bundler
Vitest ^4.0.18	Test runner
oxlint ^1.49.0	Linter (type-aware)
pnpm 10.23.0	Package manager
Node.js >= 22.12	Runtime

CLI Reference

pnpm openclaw init       # Generate config
pnpm openclaw start      # Start gateway + agent
pnpm openclaw validate   # Validate config
pnpm openclaw status     # System status
pnpm start               # Start everything
pnpm build               # Compile TypeScript
pnpm test                # Run tests
pnpm lint                # oxlint analysis

Contributing

See CONTRIBUTING.md for guidelines, maintainer list, and PR expectations. Project vision and priorities are in VISION.md.

License

MIT

Name		Name	Last commit message	Last commit date
Latest commit History 13,080 Commits
.agent/workflows		.agent/workflows
.agents		.agents
.github		.github
.pi		.pi
.vscode		.vscode
AttiClaw		AttiClaw
Swabble		Swabble
apps		apps
assets		assets
deploy		deploy
docs		docs
extensions		extensions
git-hooks		git-hooks
packages		packages
patches		patches
repos		repos
scripts		scripts
skills		skills
src		src
test		test
ui		ui
vendor/a2ui		vendor/a2ui
.detect-secrets.cfg		.detect-secrets.cfg
.dockerignore		.dockerignore
.env.example		.env.example
.gitattributes		.gitattributes
.gitignore		.gitignore
.gitmodules		.gitmodules
.markdownlint-cli2.jsonc		.markdownlint-cli2.jsonc
.npmrc		.npmrc
.oxfmtrc.jsonc		.oxfmtrc.jsonc
.oxlintrc.json		.oxlintrc.json
.pre-commit-config.yaml		.pre-commit-config.yaml
.secrets.baseline		.secrets.baseline
.shellcheckrc		.shellcheckrc
.swiftformat		.swiftformat
.swiftlint.yml		.swiftlint.yml
AGENTS.md		AGENTS.md
CHANGELOG.md		CHANGELOG.md
CLAUDE.md		CLAUDE.md
CONTRIBUTING.md		CONTRIBUTING.md
Dockerfile		Dockerfile
Dockerfile.sandbox		Dockerfile.sandbox
Dockerfile.sandbox-browser		Dockerfile.sandbox-browser
Dockerfile.sandbox-common		Dockerfile.sandbox-common
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md
VISION.md		VISION.md
appcast.xml		appcast.xml
docker-compose.yml		docker-compose.yml
docker-setup.sh		docker-setup.sh
docs.acp.md		docs.acp.md
fly.private.toml		fly.private.toml
fly.toml		fly.toml
install-ubuntu.sh		install-ubuntu.sh
install.sh		install.sh
jest.config.js		jest.config.js
openclaw.mjs		openclaw.mjs
openclaw.podman.env		openclaw.podman.env
package-lock.json		package-lock.json
package.json		package.json
pnpm-lock.yaml		pnpm-lock.yaml
pnpm-workspace.yaml		pnpm-workspace.yaml
render.yaml		render.yaml
setup-podman.sh		setup-podman.sh
setup-repos.sh		setup-repos.sh
tsconfig.json		tsconfig.json
tsconfig.plugin-sdk.dts.json		tsconfig.plugin-sdk.dts.json
tsdown.config.ts		tsdown.config.ts
vitest.config.ts		vitest.config.ts
vitest.e2e.config.ts		vitest.e2e.config.ts
vitest.extensions.config.ts		vitest.extensions.config.ts
vitest.gateway.config.ts		vitest.gateway.config.ts
vitest.live.config.ts		vitest.live.config.ts
vitest.unit.config.ts		vitest.unit.config.ts
zizmor.yml		zizmor.yml

Uh oh!

Folders and files

Latest commit

History

Repository files navigation

AttiClaw

Table of Contents

Architecture Overview

Quick Start

Installation (Ubuntu)

Manual Install

AttiClaw Web App

Technology Stack

Pages

ShadCN/UI Components

Development

Storage Architecture

Key behaviors

SMB Configuration

Model Management

HuggingFace Hub Integration

Format → Runtime Compatibility

Model Recommendations

Workflow

Upstream Feature Sources

Managing Upstream Repos

OpenClaw Core

Council System

Memory & RAG

Smart Router

Inference Backends

Skills & ClawHub

Extensions

Voice & Media

Personality & Identity

Security

Database Backends

Deployment

Proxmox Layout

Docker Compose

Systemd

Project Structure

Configuration

Testing

Build Toolchain

CLI Reference

Contributing

License

About

Resources

License

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases

Sponsor this project

Uh oh!

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages