Build in the Terminal.

AnyCLI is a command-line interface (CLI) for working with multiple LLMs directly from the terminal. It's designed for developers and teams who use the command line as a core tool and want to call AI models inside code, pipelines, IDE terminals, and CI/CD workflows.

AnyCLI provides multi-model access with explicit model and provider selection per command, including OpenAI, Anthropic, Google, Mistral, and open-source models. You always know which model you're calling and why.

AnyCLI is pipe-first and fully compatible with stdin and stdout. It works seamlessly with tools like cat, grep, jq, sed, git diff, log files and system outputs. This makes it easy to insert LLMs into existing shell workflows.

Yes. AnyCLI works inside built-in terminals of VS Code, Cursor, JetBrains IDEs, and Vim / Neovim. No plugins or extensions are required.

MCP allows AnyCLI to attach structured context and tools to model calls, including files and directories, Git repositories, external tools, and agent-based workflows. This enables advanced, context-aware LLM usage directly from the CLI.

AnyCLI provides explicit, declarative control over model and provider, temperature, max tokens, limits and runtime behavior. Nothing is hidden or auto-magical.

AnyCLI uses API keys and access policies defined at the infrastructure level. It does not abstract, proxy, or hide credentials, giving teams full control and auditability.

AnyCLI is a thin infrastructure layer between your developer workflow and LLM providers. It doesn't enforce a new workflow or hide complexity — it integrates cleanly into existing processes and scales from local scripts to CI/CD and DevOps pipelines.