VoidCMD

Your agent, autonomous by design.

[VoidCMD @VoidCMDrh]

Your models. Your rules

OpenAI, Anthropic, or any OpenAI-compatible endpoint: Ollama, LM Studio, vLLM, llama.cpp

You enter the model IDs. A connection shows Connected only after a real request succeeded

Tool calling is checked before a model runs a task with tools

[VoidCMD @VoidCMDrh]

Your keys stay on your server

Read from environment variables, or stored in your database encrypted with a key you supply

Not sent to the browser. Not written to logs, events or artifacts

The console shows the last four characters. That is all it needs

[VoidCMD @VoidCMDrh]

A task is a bounded loop, not a chat

Instruction → Plan → Execute → Checkpoint

It only goes around again while it stays inside the turns, time and budget you set

At a limit it stops with its checkpoint intact. Raise the limit, continue

[VoidCMD @VoidCMDrh]

VoidCMD. Your agent, autonomous by design

Self-hosted, on your server: a durable task runner with real tools, memory you can inspect, scheduled runs and approval gates

Bring your own keys, choose your models, resume from saved progress