L3

Llama 3.2 1B Instruct install guide

The lightest Llama — runs even on old 4 GB RAM laptops.
● Easy⏱ 10–20 min🐍 No Python⬇ ~1.3 GBFree · open source⏳ Last verified: Pending test🔒 Official sources only🧩 Skills 18
PLAYING
0:00
Video coming soon · subscribe for updatesGet updates

Install in 1 clickOFFICIAL FILES · NO REPACK · NO SIGNUP

Then run ollama run llama3.2:1b
🔒 Verified source: ollama.com⏳ Pending test
🧩 Teach this AI new skills

Unlock 9 new powers for Llama

18 skills Llama gets through Ollama, LM Studio, Open WebUI & more, mapped to what Llama can do after you install them · tap a power

🧩
Power unlocked

Llama can plug in new tools

10 skills give this power
Community Tools & Functions🟢 Official

Llama gets one-click community tools: web search, weather, code runners, filters

🌐💻🧩docs.openwebui.com · ★ 159k · in Open WebUI
Community Hub skills🟢 Official

Llama picks up community-made agent skills, prompts and slash commands

✨🧩docs.useanything.com · ★ 68.4k · in AnythingLLM
Tool calling (built-in API)🟢 Official

Llama can call your own functions: a calculator, weather, your API

💻🧩docs.ollama.com · ★ 10.7k · in Ollama
🌐
Power unlocked

Llama can search & read the web

7 skills give this power
modelcontextprotocol/servers (reference servers)🟢 Official

Llama can read your files, fetch web pages, use git, keep a memory and tell the time

📁🌐💻🧠modelcontextprotocol · ★ 92.1k · in LM Studio
Community Tools & Functions🟢 Official

Llama gets one-click community tools: web search, weather, code runners, filters

🌐💻🧩docs.openwebui.com · ★ 159k · in Open WebUI
Built-in agent skills🟢 Official

Llama can browse and scrape the web, search your docs, draw charts, save files and run SQL

🌐📁📊docs.anythingllm.com · ★ 68.4k · in AnythingLLM
📁
Power unlocked

Llama can work with your files

7 skills give this power
modelcontextprotocol/servers (reference servers)🟢 Official

Llama can read your files, fetch web pages, use git, keep a memory and tell the time

📁🌐💻🧠modelcontextprotocol · ★ 92.1k · in LM Studio
Built-in agent skills🟢 Official

Llama can browse and scrape the web, search your docs, draw charts, save files and run SQL

🌐📁📊docs.anythingllm.com · ★ 68.4k · in AnythingLLM
Jan MCP servers🟢 Official

Llama in Jan gets browser, search and file tools

🧭🌐📁docs.jan.ai · ★ 45.8k · in Jan
💻
Power unlocked

Llama can code, call tools & debug

6 skills give this power
modelcontextprotocol/servers (reference servers)🟢 Official

Llama can read your files, fetch web pages, use git, keep a memory and tell the time

📁🌐💻🧠modelcontextprotocol · ★ 92.1k · in LM Studio
lms CLI🟢 Official

Run Llama's models, local server and plugins straight from the terminal

💻lmstudio-ai · ★ 5.4k · in LM Studio
Community Tools & Functions🟢 Official

Llama gets one-click community tools: web search, weather, code runners, filters

🌐💻🧩docs.openwebui.com · ★ 159k · in Open WebUI
🧭
Power unlocked

Llama can drive a real browser

3 skills give this power
Jan MCP servers🟢 Official

Llama in Jan gets browser, search and file tools

🧭🌐📁docs.jan.ai · ★ 45.8k · in Jan
Chrome DevTools MCP🔵 Vendor-official

Llama can debug web pages in Chrome: console, network, performance traces, screenshots

🧭💻ChromeDevTools · ★ 58k · in LM Studio
Playwright MCP (browser control)🔵 Vendor-official

Llama can open web pages, click, fill forms and test sites in a real browser

🧭🌐microsoft · ★ 39.8k · in LM Studio
🔗
Power unlocked

Llama can work inside your apps

2 skills give this power
GitHub MCP server🔵 Vendor-official

Llama can search your repos, issues, PRs and Actions, and open issues and PRs for you

🔗💻github · ★ 34.5k · in LM Studio
ollmcp — MCP client for Ollama🟠 Community

Llama connects to MCP servers (files, GitHub, web) from a terminal

🧩📁🔗jonigl · ★ 836 · in Ollama
📊
Power unlocked

Llama can query data & draw charts

2 skills give this power
Built-in agent skills🟢 Official

Llama can browse and scrape the web, search your docs, draw charts, save files and run SQL

🌐📁📊docs.anythingllm.com · ★ 68.4k · in AnythingLLM
AnythingLLM agent skills🟠 Community

Llama can browse the web, work with files and query SQL in AnythingLLM

🌐📁📊Mintplex-Labs · ★ 68.4k · in Ollama
🧠
Power unlocked

Llama can remember between steps

1 skill gives this power
modelcontextprotocol/servers (reference servers)🟢 Official

Llama can read your files, fetch web pages, use git, keep a memory and tell the time

📁🌐💻🧠modelcontextprotocol · ★ 92.1k · in LM Studio
✨
Power unlocked

Llama can learn brand-new skills

1 skill gives this power
Community Hub skills🟢 Official

Llama picks up community-made agent skills, prompts and slash commands

✨🧩docs.useanything.com · ★ 68.4k · in AnythingLLM
Filters, safety notes & votes → full skills list

Step-by-step

0 / 0
  1. Pre-check: open Task Manager → Performance to see RAM and GPU. Then run this in Command Prompt to see GPU name, driver version and VRAM. Also check free disk space.
    nvidia-smi
    Do you see your GPU name and "Driver Version"?
  2. Download the Windows installer (OllamaSetup.exe, ~1.9 GB) from ollama.com/download and run it. No admin needed. Needs Windows 10 22H2+ and NVIDIA driver 551.61+.
    Do you see the Ollama icon in the system tray?
  3. Open a new Command Prompt or PowerShell and check the version.
    ollama -v
    Does it print "ollama version is …"?
  4. Run the model — the first run downloads ~1.3 GB.
    ollama run llama3.2:1b
    Do you see ">>> Send a message"? Ask anything.
  5. Check GPU use from another terminal: PROCESSOR should show "GPU". Type /bye to leave the chat.
    ollama ps
    Shows "100% GPU" or a GPU/CPU split?

Sizes are approximate and can change on the official page.

Errors & fixes

'ollama' is not recognized as an internal or external command

Open a NEW terminal after installing (old ones miss the PATH update). Reinstall if it still fails.

Error: could not connect to ollama app, is it running?

The Ollama server is not running. Windows/Mac: open the Ollama app. Linux: sudo systemctl start ollama, or run ollama serve in another terminal.

Error: model requires more system memory (… GiB) than is available (… GiB)

The model is bigger than your free RAM. Use a smaller model (e.g. llama3.2:1b) and close other apps.

Error: pull model manifest: file does not exist

Wrong model name or tag. Copy the exact name from ollama.com/library (e.g. llama3.2:3b).

ollama ps shows "100% CPU" though I have a GPU

Update the GPU driver (NVIDIA 551.61+ / latest AMD Adrenalin). If the model is bigger than VRAM, Ollama runs it on CPU — pick a smaller one.

Download stuck / very slow / "connection reset"

Network or DNS issue. Restart the router, try a mobile hotspot, or set DNS to 1.1.1.1 / 8.8.8.8. Re-running the same command usually resumes the download.

"No space left on device" / "not enough disk space"

Disk is full. Models are large — keep the free space listed above; clear old downloads and the Recycle Bin.

Your error isn't listed?

Official links

We never host or repack files. Always check the address bar shows the domain above.

Teach Llama 3.2 1B Instruct new skills 18 skills

Llama 3.2 1B Instruct is a model — it learns new skills through the app you run it in. These add-ons work with Ollama, LM Studio, Open WebUI, Jan, AnythingLLM: give the model tools, web search, files and MCP servers.

Where: Ollama, LM Studio, Open WebUI, Jan, AnythingLLM

ToolsMCPSkillsPlugins

🛡 Stay safe

Read first
  1. Official first — from the AI maker or the service’s own GitHub org / domain.
  2. Read SKILL.md and any scripts before installing. Skills can run code with your permissions.
  3. Never paste base64 or “curl … | sh” download-and-run lines from a skill.

Scan installed skills (free scanner by Snyk):

uvx snyk-agent-scan@latest <skills-folder>

Real incidents: ClawHavoc (341+ malicious skills) · Snyk ToxicSkills (76 / 3,984) · fake postmark-mcp

Community Tools & Functions

🟢 Official

One-click import of weather, web search, code runners, filters, pipes.

via Open WebUITools★ 159kLicense: see repoupdated 1 day ago
  1. openwebui.com → find tool/function → Get → enter your instance URL
  2. Review the Python code → Save
  3. Tools: Workspace → Tools; Functions: Admin Panel → Functions

⚠️ Docs: Tools/Functions execute arbitrary Python on your server; 'Featured does not mean vetted'. Only admins should import.

Can run code
Official source ↗ docs.openwebui.com⏳ Pending test Did it work?

modelcontextprotocol/servers (reference servers)

🟢 Official

Reference servers (filesystem, fetch, git, memory, time, sequential-thinking) maintained by the MCP project.

via LM Studio · Open WebUI · JanMCP★ 92.1kLicense: see repoupdated 6 days ago
e.g. npx -y @modelcontextprotocol/server-filesystem <allowed-folder>

⚠️ Reference/educational implementations; limit filesystem server to one folder.

Can run codeReads your files
Official source ↗ github.com⏳ Pending test Did it work?

Built-in agent skills

🟢 Official

Web browsing, web scraping, RAG search, chart generation, file save, SQL.

via AnythingLLMSkills★ 68.4kMITupdated 1 day ago
  1. Settings → Agent Skills → toggle skills
  2. In chat: @agent <task>

⚠️ Official. Still review what it can access before enabling.

Can run codeReads your files
Official source ↗ docs.anythingllm.com⏳ Pending test Did it work?

Community Hub skills

🟢 Official

Import community agent skills, prompts and slash commands.

via AnythingLLMTools★ 68.4kMITupdated 1 day ago
  1. Copy the item's import string → paste in AnythingLLM
  2. Docker/self-hosted: set COMMUNITY_HUB_BUNDLE_DOWNLOADS_ENABLED
  3. Review handler.js / plugin.json before enabling

⚠️ Community skills are JavaScript that runs on your machine.

Can run code
Official source ↗ docs.useanything.com⏳ Pending test Did it work?

Jan MCP servers

🟢 Official

Add browser, search and file tools to local models.

via JanMCP★ 45.8kLicense: see repoupdated 1 day ago
  1. Settings → MCP Servers → + add (command/args or URL)
  2. Enable per chat; use a tool-capable model

⚠️ MCP servers run with your permissions and can read data you give them; only connect servers you trust.

Can run codeReads your files
Official source ↗ docs.jan.ai⏳ Pending test Did it work?

Tool calling (built-in API)

🟢 Official

Give models functions (weather, calculator, your API) via /api/chat 'tools'.

via OllamaTools★ 10.7kMITupdated 2 days ago
  1. Use a tool-capable model (e.g. qwen3, llama3.1)
  2. pip install ollama → pass tools=[...] in chat()

⚠️ Your code executes the tools; validate arguments.

Can run codeActs on your account
Official source ↗ docs.ollama.com⏳ Pending test Did it work?

lms CLI

🟢 Official

Manage models, server and plugins from the terminal.

via LM StudioTools★ 5.4kMITupdated 8 days ago
npx lmstudio install-cli

⚠️ Official. Still review what it can access before enabling.

Can run code
Official source ↗ github.com⏳ Pending test Did it work?

mcpo (official MCP→OpenAPI proxy)

🟢 Official

Use any MCP server as an Open WebUI tool server.

via Open WebUIMCP★ 4.4kMITupdated 6 months ago
  1. uvx mcpo --port 8000 --api-key "top-secret" -- <your MCP server command> → Settings → Tools → add http://localhost:8000

⚠️ Use an API key; keep local.

Can run code
Official source ↗ github.com⏳ Pending test Did it work?

Pipelines (archived)

🟢 Official

Older plugin framework for custom RAG/filters in a separate server.

via Open WebUITools★ 2.4kMITupdated 20 days ago⚠ Archived — no longer updated
  1. Prefer Functions/Tools — repo archived Oct 2026

⚠️ Archived: no further fixes.

Can run code
Official source ↗ github.com⏳ Pending test Did it work?

LM Studio Hub plugins

🟢 Official

Tools-provider plugins (e.g. web tools) installable from the Hub.

via LM StudioPlugins★ 1.8kMITupdated 2 days ago
  1. Open the plugin's Hub page → Run in LM Studio (or lms clone <owner>/<plugin>)
  2. Enable via the integrations icon next to chat

⚠️ Hub plugins are user-published TypeScript; read the code.

Can run code
Official source ↗ lmstudio.ai⏳ Pending test Did it work?

Add MCP servers (mcp.json)

🟢 Official

Give local models tools via any MCP server.

via LM StudioMCP
  1. Program tab → Install → Edit mcp.json
  2. Add {"mcpServers":{"name":{"command":"npx","args":["-y","<pkg>"]}}}
  3. Enable in chat; approve tool calls

⚠️ MCP servers run with your permissions and can read data you give them; only connect servers you trust. LM Studio asks you to confirm each tool call—keep that on.

Can run code
Official source ↗ lmstudio.ai⏳ Pending test Did it work?

Chrome DevTools MCP

🔵 Vendor-official

Debug pages: console, network, performance traces, screenshots in Chrome.

via LM Studio · Jan · AnythingLLMMCP★ 58kApache-2.0updated 1 day ago

Add to LM Studio → Program tab → Install → Edit mcp.json → "mcpServers":

"chrome-devtools": {"command":"npx","args":["-y","chrome-devtools-mcp@latest"]}

⚠️ Official Chrome DevTools team. Exposes the browser's content to the AI; avoid profiles with sensitive sessions.

Can run codeReads your files
Official source ↗ github.com⏳ Pending test Did it work?

Playwright MCP (browser control)

🔵 Vendor-official

Open web pages, click, fill forms and test sites in a real browser via accessibility snapshots.

via LM Studio · Jan · AnythingLLMMCP★ 39.8kApache-2.0updated 2 days ago

Add to LM Studio → Program tab → Install → Edit mcp.json → "mcpServers":

"playwright": {"command":"npx","args":["@playwright/mcp@latest"]}

⚠️ Official Microsoft server. It can browse logged-in sites in the browser profile it uses; prefer a separate profile.

Can run codeReads your files
Official source ↗ github.com⏳ Pending test Did it work?

GitHub MCP server

🔵 Vendor-official

Read/search repos, issues, PRs, Actions; create issues and PRs on your behalf.

via LM Studio · Jan · AnythingLLMMCP★ 34.5kMITupdated 2 days ago

Add to LM Studio → Program tab → Install → Edit mcp.json → "mcpServers":

"github": {"url":"https://api.githubcopilot.com/mcp/"}

OAuth or GitHub PAT header

⚠️ Official GitHub server. Give the token/OAuth the smallest scopes you need; write actions are real.

Reads your filesActs on your account
Official source ↗ github.com⏳ Pending test Did it work?

Open WebUI (front-end with Tools)

🟠 Community

Chat UI for Ollama with Tools, Functions, MCP, RAG.

via OllamaTools★ 159kLicense: see repoupdated 1 day ago
pip install open-webui && open-webui serve

⚠️ See Open WebUI tools warning: plugins run arbitrary Python.

Can run code
Official source ↗ github.com⏳ Pending test Did it work?

AnythingLLM agent skills

🟠 Community

Use Ollama models with AnythingLLM's agent skills (web browse, files, SQL).

via OllamaTools★ 68.4kMITupdated 1 day ago
  1. Settings → LLM → Ollama
  2. Settings → Agent Skills → enable

⚠️ Review community skills' handler.js.

Can run codeReads your files
Official source ↗ github.com⏳ Pending test Did it work?

mcpo (MCP → OpenAPI proxy)

🟠 Community

Expose any MCP server as an OpenAPI tool server for Open WebUI.

via OllamaTools★ 4.4kMITupdated 6 months ago
uvx mcpo --port 8000 --api-key "top-secret" -- uvx mcp-server-time

⚠️ Set an API key; don't expose the port publicly.

Can run code
Official source ↗ github.com⏳ Pending test Did it work?

ollmcp — MCP client for Ollama

🟠 Community

Connect any MCP server (filesystem, GitHub, web) to local Ollama models in a terminal UI.

via OllamaMCP★ 836MITupdated 1 month ago
uvx ollmcp

or pip install --upgrade ollmcp

⚠️ Community tool; MCP servers run with your permissions.

Can run codeReads your filesActs on your account
Official source ↗ github.com⏳ Pending test Did it work?

Stars, license and last update come from GitHub (17 Nov 2026). Commands are copied from official docs and are not tested on a real machine yet. Access tags are our summary — always check the source. Skills for every AI →

Next guides

Alternatives All Models →

Get new free AI tools by email

Optional · one short email when this video or a new free guide lands. No spam, one-click unsubscribe.