Interactive REPL
Drop into a conversational REPL. Chat with any model, run multi-turn sessions, and use slash commands. All from your terminal.
LLMTune CLI
Inference, agents,and automation.
A developer-first CLI for running inference, coding agents, and automating LLM workflows. Straight from your command line.
Quick start
Up and running in seconds
~/projects/acme — zsh
$curl -fsSL https://llmtune.io/install.sh | bash
native binary → ~/.llmtune/bin
$llmtune login
API key saved to ~/.llmtune/config.json
$llmtune chat -p "Fix the auth test" --attach src/
[GLM-5.2] reading 4 files, working locally
Free CLIOpenAI-compatiblezero lock-in
Need an API key? Sign in to LLMTune and create one in your dashboard.
FeaturesEverything you need.
Everything you need.
Nothing you don't.
Skill system & tools
Install curated skills from the Agent Hub (llmtune skills install). Browse skills and prompts on the web at /agent.
Streaming agent loop
Run autonomous agent loops that stream reasoning, tool calls, and results in real time. Compaction and permissions built in.
OpenAI-compatible API
Built on api.llmtune.io OpenAI-compatible endpoints. Standard REST and streaming. No proprietary SDK required.
Scriptable & pipe-friendly
One-shot mode with llmtune chat -p. Pipe stdin, attach files, and chain with grep, jq, or any Unix tool in CI/CD.
Auth & key management
Run llmtune login once to store your API key in ~/.llmtune/config.json. Use llmtune doctor to verify connectivity.
Install the CLI.Start in minutes.
Install the native binary, run llmtune login with your API key, then llmtune chat to start.
macOS / Linux
curl -fsSL https://llmtune.io/install.sh | bashWindows
iex (irm https://llmtune.io/install.ps1)