runai v1.0.5

• More backends
• New settings like RUNAI_BASE_URL which lets you override the backend API base URL used for the selected model/provider. This is useful for LAN or virtual-LAN setups (such as Tailscale), local servers, LiteLLM proxies, LM Studio, Ollama, or your own OpenAI-compatible endpoint

runai v1.0.2

RunAI v1.0.2: Run or automate AI tasks from anywhere on the command line.

✨ What's new

• New backends (Gemini, Anthropic/Claude, Deepseek and more)
• Attach images/documents with -a or --append filename)
• Can now pip install . (with WIP on PyPi support)
• New env vars such as RUNAI_MODEL, RUNAI_PROJECT, RUNAI_TASK, RUNAI_TASKFILE

Example: runai -t "A short 2 line idea for helping cure aging using AI and cows" | cowsay

✨ Highlights

• Run AI tasks directly from the CLI
• STDIN / pipe support for files, source code, logs, and generated data
• Composes cleanly with Unix tools (cat, for-each, pipes, redirects)
• Pipe AI output/input like any other command.
• Supports OpenAI APIs and local models (Ollama / LM Studio, e.g. $ runai -m "ollama/deepseek-r1:8b")
• Model-agnostic, script-friendly design
• Human-readable output suitable for further piping

🧠 Philosophy

Treat AI like a Unix tool: small, composable, scriptable, and automation-first.

💼 Commercial support

Commercial support, consulting, and custom automation available — DM or get in touch.

runai v1.0.0

runai: Run or automate AI tasks from anywhere on the command line.

✨ Highlights

• Run AI tasks directly from the CLI
• STDIN / pipe support for files, source code, logs, and generated data
• Composes cleanly with Unix tools (cat, for-each, pipes, redirects)
• Pipe AI output/input like any other command.
• Supports OpenAI APIs and local models (Ollama / LM Studio)
• Model-agnostic, script-friendly design
• Human-readable output suitable for further piping

🧠 Philosophy

Treat AI like a Unix tool: small, composable, scriptable, and automation-first.

💼 Commercial support

Commercial support, consulting, and custom automation available — DM or get in touch.

v0.8.0-beta

Athough still beta, a lot of functionality works