AITerm is a cross-platform AI-powered terminal. Beyond cloud models, it runs local models to save on token costs and keep your data on-box — or bridges the Claude Code CLI to any provider you want.
More than that, it treats what the AI produces as work: a task board lets you queue jobs, and AITerm dispatches them to Claude Code, tracks the outcome, and writes them up as a report.
Autonomous agent loops, multi-database connections, a knowledge base, and MCP tools are all built in, on Tauri 2, React 19, and Rust.
Models can already write code, run commands, and carry a multi-step task to the end. You, meanwhile,
still watch one thing at a time: give the instruction, wait, read the result, start the next.
The bottleneck stopped being what the AI can do. It is how many things you can personally supervise.
Running ten AI tasks at once is technically trivial. What is hard is remembering which one is running, which finished, and which failed and needs another go. The missing piece is not a better model — it is somewhere to put all of that.
The same task, stated precisely or vaguely, produces wildly different results. Rather than correcting after the fact, write down the goal, the scope, the constraints and what "done" means up front — which is also what makes a piece of work handed-off-able at all.
Automation that evaporates when it finishes accumulates nothing. What was done, how, and how it turned out has to be reviewable afterwards and presentable as a report. That is what makes it work rather than just an execution.
Write what you want done as a card: a title, which folder to work in, detailed instructions, attachments if you need them. Drag it to Queued and AITerm opens a tab, launches Claude Code, delivers the instructions, and moves the card to Done with a success or failure outcome.
Line up an evening's work and read the results later, instead of driving each task by hand. How many run at once is yours to set — nothing gets fired off all at once.
Turn a project's board into a document. The AI reads every card — including the full transcripts of finished work — and writes a formatted report saved into the project folder as history. A finished card's summary is written once and cached, so the next report only processes what is new.
Jot a rough note into a task, press Refine, and the AI rewrites it into something executable: goal, scope, constraints, acceptance criteria — without inventing work you never asked for. One click restores your original.
Finished work does not pile up forever. Archive one card or a whole column and the board keeps only what is in flight. Archiving is not deleting — cards, transcripts and attachments stay put, searchable, pageable, and restorable to the board.
AI built directly into your workflow — no window switching required
Type /ai <query> in the terminal to instantly generate a shell command with a risk level shown — confirm and run with one click.
Beyond cloud providers, AITerm connects to Ollama, LM Studio, oMLX, and other local / OpenAI-compatible endpoints — inference runs on your own machine with no API cost and nothing sent off-box, which pays off for high-volume automation. Or use the Claude Code Bridge to run the Claude Code CLI on any provider you have configured.
Every command and its output renders as its own card instead of raw scrollback. The live terminal pane auto-expands while a command is running and shrinks back down when idle, keeping the screen clean.
Type /agent to start multi-step agent mode — the AI plans and executes complex tasks on its own, falling back to manual confirmation for risky actions.
"Loop engineering" lets the AI break down and execute tasks on its own — no back-and-forth confirmation needed. An orchestrator keeps dispatching sub-agents until the goal is complete, with safety gates and a live execution trace watching every step.
Queue work as cards and AITerm opens a tab, launches Claude Code, delivers the instructions, and marks the outcome. Includes projects, interactive mode, transcripts, AI work reports, and an archive. See more →
A dedicated tab where the AI answers questions about any project directory — it scans the file tree, searches and reads source files on its own, can render Mermaid diagrams, and lets you export the conversation to Markdown.
A context-aware AI chat panel built right next to the terminal — for troubleshooting, code generation, and brainstorming, all in one place.
Reports, comparison tables, and charts the AI produces open in a formatted panel beside the conversation instead of raw markup in a chat bubble. Bar, line, and pie charts render directly, switch to a table view for exact numbers, and save to an HTML file in one click. Available across the terminal and remote AI panels, Code Assistant, Knowledge Base, and database / cross-database Q&A.
Create a notebook pointed at a folder and AITerm chunks, embeds, and indexes every document inside it, so you can ask natural-language questions with answers that cite the source documents. Works with Ollama or any OpenAI-compatible embedding provider.
Native support for PostgreSQL, MySQL, SQLite, and ODBC. DB2 works on both macOS and Windows via a built-in Java JDBC sidecar — no driver installation required.
A built-in sidebar file browser kept in two-way sync with the terminal's working directory — navigate in the terminal and the explorer follows, or click the "cd" button (or the folder picker built into the command input) to jump the terminal to any folder instantly.
Ctrl+F searches across your entire block history with live scroll-to-match; Ctrl+Shift+R opens a bookmarks picker to save and refill frequently used commands.
Full support for English and Traditional Chinese, switchable at runtime with no restart required.
Works out of the box with OpenAI, Anthropic, Google, Ollama (local inference), and OpenAI-compatible endpoints — switch providers whenever you need to. You can also sign in with a subscription account — Claude, ChatGPT, Gemini, or GitHub Copilot — with no API key to obtain.
Connect any MCP (Model Context Protocol) server and the AI can read and write files, query data, and call external services directly — no terminal command in between. Every tool call and its result stays visible and expandable.
Point the Claude Code CLI at any provider you have configured in AITerm — run Claude Code on your OpenAI, Gemini, or local models, and switch between them at any time.
Once you connect to a colleague's shared terminal, Ask AI is the same full chat panel as local — free-form Q&A, provider switching, and browsing or continuing past conversations, all through your own configured provider, so the other machine needs no AI setup. Run /agent in a remote tab to have the AI plan, issue commands, and read results step by step.
Enable the Remote toggle to send commands to your terminal from your phone via a Telegram bot — results are sent right back to the same Telegram chat, so you're never far from your session.
AI-assisted feature discussions that automatically generate spec documents (SDD) into your project's structured directory — turning ideas into trackable specs.
A modern desktop app architecture connecting a high-performance Rust backend to a React frontend via Tauri IPC
.dmg.dmg and drag AITerm into the "Applications" folderxattr -cr /Applications/AITerm.app
-setup.exelibwebkit2gtk-4.1). Ubuntu 20.04 and earlier are not supported.
_x86_64.AppImage or _amd64.debchmod +x AITerm_*_x86_64.AppImage && ./AITerm_*_x86_64.AppImage
sudo apt install ./aiterm_*_amd64.deb
@BotFather on Telegram, send /newbot to create a new bot, and get your Bot Tokenhttps://api.telegram.org/bot<YOUR_TOKEN>/getUpdates
📱 Remote toggle in the terminal status bar to enable remote controlRequires Rust 1.78+, Node.js 20+
git clone https://github.com/jamesju9999/AITERM.git cd AITERM npm install npm run tauri:dev