safeclaw

# SafeClaw 🐾

**The zero-cost alternative to OpenClaw. No LLM required, though it is optional. No required API bills. Minimal attack surface. Runs on any machine.**

> **ONE HUNDRED STARS** — We hit ONE HUNDRED stars on GitHub! This milestone is our biggest update since our original release. It brings: **Real Research, Simple AI Setup, and Smart Learning.** Research now searches arXiv and Semantic Scholar for real academic papers, asks Wolfram Alpha for computational answers. `setup ai sk-ant-your-key` instantly configures Anthropic (or any provider) — or `setup ai local` for free Ollama. The parser auto-corrects typos, converts word-numbers ("one" → 1), and learns from your mistakes automatically. [See the new features below.](#-real-research-sources) Previous milestone: [Fuzzy Learning & Personalization.](#-writing-style-profiler)
While OpenClaw users are burning [$200/day](https://www.notebookcheck.net/Free-to-use-AI-tool-can-burn-through-hundreds-of-Dollars-per-day-OpenClaw-has-absurdly-high-token-use.1219925.0.html) and [$3,600/month](https://dev.to/thegdsks/i-tried-the-free-ai-agent-with-124k-github-stars-heres-my-500-reality-check-2885) on API tokens, SafeClaw delivers 90% of the functionality using traditional programming — rule-based parsing, ML pipelines, and local-first tools. **Your API bill: $0. Forever.** If you choose to use the language models for blogging, coding and research, then yeah, you'll have a bill, but you can also do all of that without a language model on safeclaw. We simply provide both options. 

SafeClaw uses VADER, spaCy, sumy, YOLO, Whisper, Piper, and other battle-tested ML techniques instead of generative AI. The result: deterministic, predictable, private, and completely free to run.

---

## Why SafeClaw?

| | SafeClaw | OpenClaw |
|---|---|---|
| **Monthly cost by default** | **$0** | $100–$3,600+ | 
| **Requires LLM** | No (optional for AI blogging, coding, and research all have no LLM options) | yes 
| **Prompt injection risk** | **Minimal** (only if a url you go to is malicious) | Yes (everywhere) |
| **Works offline** | **Yes** all default features, except weather | No |
| **Runs on any machine** | **Yes** (Linux, macOS, Windows) | Needs powerful hardware or cloud APIs |
| **Deterministic output** | **Yes** | No (LLM responses vary) |
| **Privacy** | **Local by default** (external only when you ask or setup an llm, e.g. weather) | Data always sent to API providers |


## Full Comparison: SafeClaw vs OpenClaw

| Feature | SafeClaw | OpenClaw |
|---|---|---|
| Self-hosted | ✅ | ✅ |
| Cross-platform (Linux, macOS, Windows) | ✅ | ✅ |
| No AI/LLM required | ✅ | ❌ |
| Offline capable | ✅ | ❌ |
| Zero API cost | ✅ | ❌ |
| Prompt injection risk | **Minimal** (only LLM research) | Everywhere |
| Privacy-first | ✅ (local by default) | ✅ |
| Voice (STT/TTS) | ✅ (Whisper + Piper, local) | ✅ (ElevenLabs, paid API) |
| Smart home control | ✅ | ✅ (via skills) |
| Bluetooth control | ✅ | ❌ |
| Network scanning | ✅ | ❌ |
| Social media summaries | ✅ (Twitter, Mastodon, Bluesky) | ❌ (requires separate skills) |
| Multi-channel | ✅ (CLI, Telegram, Webhooks) | ✅ (13+ platforms) |
| Web crawling | ✅ | ✅ |
| Summarization | ✅ (extractive) | ✅ (AI-generated) |
| RSS/News feeds | ✅ (50+ feeds) | ✅ (via skills) |
| Sentiment analysis | ✅ (VADER) | ✅ (AI) |
| Email integration | ✅ | ✅ |
| Calendar support | ✅ | ✅ |
| Document reading | ✅ | ✅ |
| Desktop notifications | ✅ | ✅ |
| Object detection | ✅ (YOLO) | ❌ |
| OCR | ✅ (Tesseract) | ❌ |
| Cron jobs | ✅ | ✅ |
| Webhooks | ✅ | ✅ |
| Plugin system | ✅ | ✅ (5,700+ skills) |
| Free-form chat | ❌ | ✅ | 
| Blog with and without  LLM  | ✅ no llm (extractive titles) with LLM, AI blogging | ❌ (always requires AI) |   
Blog publishing (WordPress, Joomla, SFTP) | ✅ | ❌ (requires plugins) |
| Writing style learning | ✅ (statistical profiling) | ❌ |
| Academic research (arXiv, Scholar) | ✅ (free, no API key) | ❌ |
| Wolfram Alpha | ✅ (computational knowledge) | ❌ |
| Simple  setup (enter key or local) simple telegram and blogging setup | ✅ (one command) | ❌ |
| Auto-learning from mistakes | ✅ (word-to-number, typo correction) | ❌ |
| Research pipeline | ✅ (two-phase, LLM optional) | ✅ (AI only) |
| Code templates & tools | ✅ (7 templates, offline utils) | ❌ (requires AI) |
| Auto-blog scheduling | ✅ (cron-based) | ❌ |
| Task-aware prompts | ✅ (per-task LLM routing) | ✅ |
| Command chaining | ✅ ("read email and remind me at 3pm") | ✅ |
| Autonomous multi-step tasks | ❌ | ✅ |
| Self-writing skills | ❌ | ✅ |
| Browser automation | ❌ | ✅ |

---

## Features

### 🗣️ Voice Control
* **Speech-to-Text** — Whisper STT runs locally, no cloud transcription
* **Text-to-Speech** — Piper TTS for natural voice output, completely offline
* **Voice-first workflow** — Talk to SafeClaw like you would any assistant

### 🏠 Smart Home & Device Control
* **Smart home integration** — Control your connected devices
* **Bluetooth device control** — Discover and manage Bluetooth devices
* **Network scanning** — Device discovery on your local network

### 📱 Social Media Intelligence
* **Twitter/X summarization** — Add accounts, get summaries of their activity
* **Mastodon summarization** — Follow and summarize fediverse accounts
* **Bluesky summarization** — Track and summarize Bluesky feeds
* No API tokens needed for public content

### 📰 RSS News Aggregation
* **50+ preset feeds** — Hacker News, Ars Technica, BBC, Reuters, Nature, and more
* **8 categories** — Tech, World, Science, Business, Programming, Security, Linux, AI
* **Custom feeds** — Import any RSS/Atom feed
* **Auto-summarization** — Extractive summaries with sumy (no AI)
* **Per-user preferences** — Customize your news sources

### 🔒 Privacy & Security
* **Self-hosted by default** — Your data stays local unless you explicitly request external info (like weather)
* **No API keys required** — Core features work completely offline
* **No cloud AI dependencies** — No tokens sent to OpenAI, Anthropic, or Google
* **Minimal prompt injection surface** — Core features (blogging, coding, commands) use rule-based parsing with zero LLM, so they're immune to prompt injection. Research and deep analysis *can* optionally use an LLM to synthesize sources — when enabled, content from external sources (arXiv abstracts, web pages) is fed to the LLM, which carries a minimal prompt injection risk. Phase 1 research (gathering/summarizing) never uses an LLM and is safe. We're transparent about this tradeoff.

### 📡 Multi-Channel
* **CLI** — Interactive command line with Rich formatting
* **Telegram** — Full bot integration
* **Discord** — Coming soon
* **Slack** — Coming soon
* **Webhooks** — Inbound and outbound support

### ⚡ Automation
* **Command chaining** — Combine actions naturally: "read my email and remind me at 3pm"
* **Web crawling** — Async crawling with depth limits and domain filtering
* **Summarization** — LexRank, TextRank, LSA, Luhn algorithms
* **Reminders** — Natural language time parsing with dateparser
* **Shell commands** — Sandboxed command execution
* **File operations** — Search, list, read files
* **Cron jobs** — Scheduled task automation
* **Daily briefings** — Weather, reminders, news from your feeds

### 📊 Text Analysis
* **VADER Sentiment** — Lexicon-based sentiment analysis
* **Keyword Extraction** — TF-IDF style extraction
* **Readability Scoring** — Flesch-Kincaid metrics

### 📧 Email Integration
* **IMAP Support** — Read emails from Gmail, Outlook, Yahoo
* **SMTP Support** — Send emails
* **Standard protocols** — No API keys required

### 📅 Calendar Support
* **ICS Files** — Import and parse .ics calendar files
* **CalDAV** — Connect to Google Calendar, iCloud (optional)
* **Event filtering** — Today, upcoming, by date range

### 📄 Document Reading
* **PDF** — Text extraction with PyMuPDF
* **DOCX*

[truncated…]