GitHub - siddsachar/Thoth: Thoth - Personal AI Sovereignty. A local-first AI assistant with integrated tools, a personal knowledge graph, voice, vision, shell, browser automation, scheduled tasks, health tracking, and messaging channels. Run locally via Ollama or add opt-in cloud models. Your data stays on your machine.
Pangram verdict · v3.3
We believe that this document is a mix of AI-generated, and human-written content
AI likelihood · overall
MixedArticle text · 1,408 words · 5 segments analyzed
Thoth is a local-first AI assistant for personal AI sovereignty: a desktop agent with memory, tools, workflows, design creation, messaging, plugins, and optional cloud models while your durable data stays on your machine. It runs fully local through Ollama with 39 curated tool-calling models, or you can opt into OpenAI, Anthropic, Google AI, xAI, OpenRouter, and ChatGPT / Codex when you want frontier reasoning or do not have a GPU. API keys and in-app subscription tokens are stored in the OS credential store when available; Thoth has no account system, server, or telemetry pipeline.
🖥️ One-click install on Windows & macOS — download, run, done. No terminal, Docker, or config files required. Get it here.
Why the name "Thoth"? In ancient Egyptian mythology, Thoth (𓁟) was the god of wisdom, writing, and knowledge — the divine scribe who recorded all human understanding. Like its namesake, this tool is built to gather, organize, and faithfully retrieve knowledge — while keeping everything under your control.
✨ Why Thoth Stands Out
Full subsystem details live in docs/ARCHITECTURE.md.
Memory That Becomes A Knowledge Graph Thoth stores durable knowledge as entities and typed relationships, not just chat snippets. It can save, search, link, explore, visualize, and export your knowledge graph as an Obsidian-compatible wiki vault, while background extraction and Dream Cycle refine duplicates, stale confidence, missing relationships, and actionable insights. Designer Studio, Media, And Documents Designer Studio creates decks, documents, landing pages, app mockups, and storyboards with a sandboxed interactive runtime, critique-repair loop, editable exports, and shareable HTML. The same media layer supports image generation/editing with OpenAI, Google, and xAI, plus video generation through Google Veo and xAI Grok Imagine Video. Real Computer Use, With Guardrails The LangGraph ReAct agent has 30 core tool modules plus auto-generated channel tools. It can browse in a visible Chromium window, use the shell, manage files, search the web, read documents, send email, manage calendar events, create charts, inspect system status, and call external MCP tools.
Destructive actions are approval-gated, shell commands are safety-classified, the filesystem is sandboxed, and tools can be disabled individually. Workflows That Run On Their Own Advanced workflows add schedules, webhook triggers, condition branches, approval steps, subtasks, notification-only runs, concurrency groups, per-workflow model/tool/skill overrides, and background safety modes. The Workflow Console shows running jobs, approvals, recent history, and insights without leaving chat. Native, Multi-Channel, Local-First Thoth ships as a desktop app for Windows and macOS with one-click installers, tray integration, setup wizard, voice input, local Kokoro TTS, desktop notifications, and 5 bundled messaging channels: Telegram, WhatsApp, Discord, Slack, and SMS. Channels share media handling, streaming, approvals, health checks, and optional tunnel support. Extensible Without Giving Up Control Plugins add tools and skills through a sandboxed marketplace; MCP servers add external tools with per-server and per-tool review; Claude Code Delegation can coordinate Claude Code CLI as an approval-gated external coding worker; the migration wizard imports selected Hermes/OpenClaw data with backups and redacted reports. Core and plugin API keys use the OS credential store when available, with metadata-only files in Thoth's data directory. Feature Map
Area What Thoth Includes
Agent & Models LangGraph ReAct agent, streaming, thinking bubbles, smart context trimming, 39 curated Ollama models, opt-in OpenAI/Anthropic/Google/xAI/OpenRouter provider models plus ChatGPT / Codex subscription models, per-thread and per-workflow model overrides
Memory & Knowledge Personal knowledge graph, FAISS semantic recall, 67 typed relations, graph visualization, Obsidian wiki export, document extraction, Dream Cycle refinement, self-knowledge, and insights
Design & Media Designer Studio, PDFs/HTML/PNG/PPTX export, published interactive links, image generation/editing, video generation, chart insertion, Mermaid/Plotly rendering, and media persistence
Tools 30 core tools covering search, browser, shell, filesystem, documents, Gmail, Calendar,
X, memory, workflows, tracker, image/video, vision, status, MCP, updates, computation, weather, charts, and system info
Automation Scheduled workflows, step pipelines, conditions, approvals, subtasks, webhooks, task-completion triggers, notifications, channel delivery, run history, and safety modes
Channels & Voice Telegram, WhatsApp, Discord, Slack, SMS, local faster-whisper STT, Kokoro TTS, media intake, reactions, streaming, approval routing, and tunnel manager
Platform & Extensibility Native desktop app, one-click installers, auto-updates, plugin marketplace, MCP client, migration wizard, configurable identity, secure API-key storage, 13 manual skills including Claude Code Delegation, and 18 tool guides
Detailed architecture and subsystem reference →
How does Thoth compare to OpenClaw? OpenClaw is a popular open-source personal AI assistant. It's a powerful multi-channel gateway built for developers comfortable in the terminal. Here's how the two compare:
Thoth OpenClaw
Getting started One-click installer (.exe / .dmg) — download, run, done. Built-in setup wizard, no terminal required npm install -g openclaw@latest → CLI onboarding. Requires Node.js 24. Windows needs WSL2 (no native Windows support)
Local AI (offline) Local-first — Ollama with 39 curated models out of the box. Works fully offline. Provider models are opt-in Cloud-first design — requires an API key to start. Local model support through provider config
Memory Personal knowledge graph — 10 entity types, typed directional relations, visual explorer, FAISS semantic search + 1-hop graph expansion, memory decay, orphan repair Flat markdown files (MEMORY.md + daily notes) with semantic search. No structured graph
Knowledge refinement Dream Cycle — 5-phase nightly engine: duplicate merging (≥0.93 similarity), description enrichment, stale-confidence decay, relationship inference with hub diversity caps and rejection cache, and actionable insight generation.
3-layer anti-contamination system, dream journal Dreaming (experimental) — Light/Deep/REM phases that promote short-term signals to long-term memory via scoring thresholds
Document intelligence Map-reduce LLM pipeline — extracts structured entities and relations into the knowledge graph with source provenance. Curated 67-type relation vocabulary, entity caps, self-loop rejection. Supports PDF, DOCX, EPUB, HTML, Markdown File read/write/edit operations in the workspace
Wiki vault Obsidian-compatible export — one .md per entity with [[wiki-links]], YAML frontmatter, and per-type indexes Not available
Voice Fully local — faster-whisper STT + Kokoro TTS with 10 voices. Audio never leaves your machine ElevenLabs (cloud TTS) + system fallback. Voice Wake on macOS/iOS
Health tracking Built-in tracker — medications, symptoms, exercise, mood, sleep, periods. Streak analysis, CSV export, Plotly charts Not available
Tools 30 core tools plus auto-generated channel send tools — shell, browser, filesystem, Gmail, Calendar, Designer Studio, Thoth Status, memory graph, MCP external tools, image + video generation, and research tools ~20 built-in tools — exec, browser, web search, canvas, cron, image/music/video generation
Messaging channels 5 channels — Telegram, WhatsApp, Discord, Slack, SMS — all with streaming, reactions, media, and approval routing. Auto-generated per-channel tools. Tunnel manager for webhooks 23+ channels — WhatsApp, Telegram, Slack, Discord, Signal, iMessage, Teams, Matrix, IRC, and many more
Autonomous agents Advanced workflows — step-based pipelines with conditions, approval gates, webhook triggers, concurrency groups, and per-workflow safety mode. Multiple run in parallel with their own persistent threads Multi-agent routing with isolated sessions per sender/channel
Desktop app Native window (pywebview) + system tray on Windows & macOS. One-click installers for both macOS menu bar app. No native Windows app (WSL2 required).
iOS & Android companion apps
Designer / Canvas Designer Studio for decks, one-pagers, reports, published links, plus Mermaid diagrams and Plotly charts rendered inline A2UI — agent-driven interactive visual workspace
Plugins Sandboxed plugin marketplace with hot-reload and security scanning npm plugin ecosystem + ClawHub skill registry. Large community catalog
Privacy All data local. No account, no server, no telemetry. API keys use the OS credential store when available — Thoth has no servers Self-hosted gateway. Data stays on your machine. Some channel integrations require external services
Cost Free with local models. Provider models use upstream API billing or ChatGPT subscription access only when you opt in Free + open source. Requires a cloud API key to function
In short: OpenClaw is a powerful gateway for developers who want their AI assistant on every messaging platform. Thoth is built for people who want personal AI sovereignty — local-first intelligence, a structured knowledge graph that grows with you, one-click setup, and tools that work without touching a terminal. Different philosophies, both open source.
For comparisons with ChatGPT and other cloud assistants, see docs/ARCHITECTURE.md.
🔧 Tools Thoth's agent has access to 30 core tool modules. Many of them expose multiple operations, and running messaging channels add extra send/photo/document tools automatically. Tools can be enabled or disabled from the Settings panel. Search & Knowledge
Tool Description API Key?
🔍 Web Search Live web search via Tavily for current events, news, real-time data TAVILY_API_KEY
🦆 DuckDuckGo Free web search — no API key needed None
🌐 Wikipedia Encyclopedic knowledge with contextual compression None
📚 Arxiv Academic paper search — newest-first sorting, full-text HTML links, arXiv query syntax (ti:, au:, abs:, cat:) None
▶️