Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/jorgeurtubiam-ship-it/Gulin_ia/llms.txt

Use this file to discover all available pages before exploring further.

This guide takes you from zero to your first AI-powered command in GuLiN Terminal. By the end you will have the application running locally, an AI provider configured (either a free local model via Ollama or a cloud provider of your choice), and hands-on experience with GuLiN’s core features: AI command completion, the AI chat panel, and hot model switching.
1

Install GuLiN Terminal

Follow the Installation guide to build GuLiN from source, or download a pre-built installer for your platform from gulin.dev/download.Once you have cloned the repository and run task init, you are ready to launch.
2

Launch the application

Start GuLiN in development mode from the root of the repository:
task dev
The GuLiN Terminal window will open automatically. The Go backend, Electron shell, and Vite dev server all start together. Hot Module Reloading is active, so any frontend changes you make are reflected immediately.
3

Choose an AI provider

Click the model dropdown in the top-left corner of the GuLiN window. You will see all available providers. Choose the one that fits your setup:
Ollama runs large language models entirely on your own machine — no API key and no data leaving your device.
  1. Make sure Ollama is installed and running: ollama.com
  2. Pull a model. Llama 3.2 is a great starting point: 
    ollama run llama3.2
  3. In the GuLiN model dropdown, select Local (Ollama) or Custom and confirm the endpoint is set to: 
    http://localhost:11434
GuLiN also uses Ollama’s nomic-embed-text model for its Gulin Brain memory embeddings. Run ollama pull nomic-embed-text to enable fully local, private long-term memory.
You can hot-switch models at any time — including mid-conversation. GuLiN shares its chat history and Gulin Brain memory across all providers, so switching from Ollama to GPT-4o (or back) never resets your context or loses your work.
4

Open a terminal block and try AI command completion

GuLiN’s tiled workspace starts with a terminal block open by default. Click inside the terminal to focus it.Type a natural-language comment starting with # and press TAB:
# check disk space
After pressing TAB, GuLiN’s AI will predict and insert the appropriate shell command — for example df -h — directly into your prompt. Press Enter to run it.
The # + TAB shortcut works on any open terminal block. The AI reads the current working directory, recent output, and your Gulin Brain memory to predict the most relevant command for your context. No chat window interaction required.
Try a few more examples to get a feel for the feature:
# list all running docker containers

# find files larger than 100MB in the current directory

# show memory usage by process
5

Ask GuLiN AI for help in the AI panel

Click the AI panel (the chat icon on the left sidebar, or open a new AI block from the workspace menu). This is GuLiN’s full agentic assistant — it can read your terminal output, manage files, query databases, browse the web, and execute commands with your approval.Try asking it something that requires multi-step reasoning:
Analyze the disk usage in my home directory and tell me the top 5 largest directories.
GuLiN will:
  1. Use the term_run_command tool to run du -sh ~/* (or equivalent) in your terminal
  2. Read and summarize the output
  3. Present a ranked summary in the chat
You will see each tool call appear in the chat as a collapsible step — you can inspect exactly what the AI did and why.
All agentic actions that touch your system (running commands, clicking web elements, writing files) require your explicit approval in the chat. GuLiN never executes anything autonomously without confirmation.
6

Explore the tiled workspace

GuLiN’s workspace supports multiple block types that you can arrange freely:
  • Terminal block — standard shell with AI command completion
  • AI panel — the full agentic assistant
  • File editor — syntax-highlighted editor for local and remote files
  • Web browser — agentic browser the AI can navigate on your behalf
  • Dashboard — charts and data grids rendered inline
Drag any block header to reposition it. Use the + button to open additional blocks. All panels share the same GuLiN Brain context, so the AI assistant is always aware of what is happening across your entire workspace.

​
What’s Next

Now that GuLiN is running, explore some of its more powerful features:
  • Gulin Brain — Review and extend the AI’s long-term memory using brain_list and brain_update in the AI panel
  • DB Maestro — Connect to a database (Oracle, Postgres, MySQL, MongoDB, SQLite, or Dremio) from the workspace menu and let the AI write and run queries for you
  • SSH Sessions — Open an SSH block and connect to a remote server; the session will survive network interruptions and reconnect automatically
  • Interactive Dashboards — Ask the AI to visualize any data as a bar, line, area, pie, or radar chart directly in the workspace
  • GulinApp SDK — Build custom terminal micro-applications using the Tsunami framework (Go + VDom)

Build docs developers (and LLMs) love

Get started for freeTalk to us