LMIM OS – an offline AI ecosystem. Voice, RAG, WhatsApp. ++ One file. 0 setup

iamonthemission1 pts0 comments

LMIM OS v2.1 'Tezcat · Sharpened' — Linux + Windows Live Now

What can LMIM do?<br>19+

Everything. On your machine.

19+ tools — no cloud, no API key, no subscription. All in one AppImage / Installer.

⚡ Core AI

🧠Local LLMQwen 3.5 · bundled

📄RAG LiteDocument Q&A · local embeddings

📁WorkspaceSandboxed file ops

🎙VoiceSTT + TTS · 5 languages

💭Persistent memoryEpisodic + semantic

⚡CUDA acceleration3–15× faster on NVIDIA

🎯Prime DirectiveStanding instructions

🛠 Operations

🤖Build systemPlanner · Builder · Inspector

🚀Campaign BlasterBulk WhatsApp + email

🧰Developer toolboxHash · minify · JSON

📅Visual agendaCalendar + scheduling agent

📡 Services

💬WhatsAppBaileys · QR · auto-reply

✈️TelegramBot API · polling

📧EmailSMTP · IMAP

💼SlackBot · webhooks

🎮DiscordBot · webhooks

🧬 Intelligence

🕷Web Scraper10 URLs · batch · LMIM analysis

📇ContactsNatural name resolution

🎨Image generationOn-device · v2.2

🛡Horus securityAudit trails · v3.0

⬇ Download<br>Learn more

⬢ Tezcat · Sharpened · v2.1.1

LMIM OS v2.1 — The same fire. Sharper edge.

RAG. Workspace sandbox. Web scraper. Contacts. Voice on GPU. Bundled model. CUDA. Linux + Windows live now.

📄 RAG Lite — doc Q&A

📁 Workspace sandbox

🕷 Web Scraper · batch

📇 Contacts · names

✦ Also in v2.1 / v2.1.1

🎯 Prime Directive — standing instructions<br>🧠 Hardened 4-layer tool parser<br>🎙 Voice device routing + CPU bug fixes

Download v2.1 — Linux + Windows →

Download manual (.md)

Shape v3 — what should we build?

🐧 Linux · live now<br>🪟 Windows · live now<br>$0 forever

What should we focus on? (tap any · multi-select)

🧠 More models / model variety<br>⚡ Speed & performance<br>🔗 New integrations / daemons<br>🎨 UI / UX polish<br>🎙 Voice & audio<br>🤖 Smarter agents / planning<br>🧰 More developer tools<br>📱 Mobile / remote access<br>📚 Documentation & tutorials<br>🐛 Bug fixes & stability<br>🏢 Enterprise / team features<br>✨ Something else

Tell us more (optional · the more specific the better)

Want a follow-up? (optional · only used to reply)

Maybe later

Send to LMIM →<br>Sending…

Stored locally on the LMIM server as JSON. Email used only to reply. No tracking.<br>Privacy policy.

Logged. Thank you.

Your input goes straight to the v3 planning notes. If you left an email, expect a reply within a week.

Close

⬢ v2.1 'Tezcat · Sharpened' is live now — RAG, workspace, scraper, contacts, CUDA, bundled model. Linux + Windows Live · Download v2.1 →

⬢ v2.1 — Tezcat · Sharpened<br>🐧 Linux Live Now · 🪟 Windows Live Now<br>✦ $0 forever · model included

LEAN MEAN INFERENCE MACHINE

Local AI That

Actually Does Things

The same fire. Sharper edge. RAG, sandboxed workspace, web scraper, contacts, voice on GPU.

Bundled model. CUDA acceleration. One file. Everything included. Linux + Windows live now.

Download v2.1 — Live Now

See what's new →

LMIM OS v2.1 · Tezcat · Sharpened · Linux + Windows Live

1,000+ downloads

20+ countries

Zero cloud / telemetry

MIT open source

$0 forever

Scroll to explore

⬢ v2.1 'Tezcat · Sharpened' — Live Now · Linux AppImage + Windows Installer

RAG Lite: Drop a PDF/TXT/MD — ask anything about it. Local embeddings, never leaves your machine.

Workspace sandbox: Point LMIM at a folder. Reads, writes, creates, edits. Path traversal blocked at backend.

Web Scraper: Batch 10 URLs in parallel. LMIM analyzes them based on your stated purpose.

Contacts: Natural-language name resolution — say "Send WhatsApp to Maria" and she's found.

4-layer tool parser: Strict → embedded → relaxed → regex. No more silent failures mid-chain.

CPU/GPU toggle + one-click model download: Qwen 3.5 9B streamed straight to ~/.lmim_os/models/.

Windows CPU Voice Fix: Solved audio routing bugs on CPU. Whisper STT and Piper TTS now run flawlessly on Windows without GPU.

Chat Performance: Optimized token streaming and UI rendering for smoother, faster conversations.

✦ v3 in planning — Horus security engine, image generation, plugin system. See what's coming →

📄

RAG Lite + Workspace

Drop a document. Ask anything about it. Point it at a folder and let it build.

Local embeddings · all-MiniLM-L6-v2

Smart chunking · respects headings + code blocks

MMR reranking · non-redundant retrieval

Sandboxed file ops · safe_path() blocks traversal

Just say it.

"Summarize this contract in 3 bullets."<br>Reads the PDF, answers only from what's inside.

"Refactor main.py to use async."<br>Reads, edits, runs — entirely within your workspace.

"What are the risks in this doc?"<br>Suggested prompts auto-generated on upload.

🕷

Scraper + Contacts

Batch 10 URLs in parallel. Name-resolve anyone in your contacts.

Parallel fetch · 30s per URL

Basic mode or LMIM analysis mode

Natural-language contact lookup

"Send WhatsApp to Maria" — resolved automatically

Token generation · Qwen 3.5 2B<br>tok/s

CPU only

17

GTX 1650 Ti

80

RTX 3060

120

measured · same model · same prompt

Hardened.

4-layer tool parser.<br>Strict → embedded → relaxed → regex. No more silent failures.

Qwen3 reasoning_content...

live lmim windows linux workspace contacts

Related Articles