Deep-ai-inc/ch.at

Universal Basic Chat

What it solves

It provides a lightweight, privacy-focused way to access large language models (LLMs) across multiple network protocols without requiring a web browser, JavaScript, or user accounts.

How it works

Written in Go as a single static binary, the service acts as a gateway to LLM providers like OpenAI, Anthropic, or local models via Ollama. It exposes a chat interface through four different methods: HTTP (web/curl), SSH, DNS (via TXT records), and an OpenAI-compatible API.

Who it’s for

Developers and power users who want minimal, low-overhead access to AI via the terminal or network tools, and those who prioritize a "no-tracking" architecture.

Highlights

  • Multi-protocol access: Interact with AI via curl, ssh, dig (DNS), or standard API calls.
  • Privacy-centric: No server-side logs, no user accounts, and no tracking; history is stored only on the client side.
  • Ultra-lightweight: Approximately 1,300 lines of Go code with only three direct dependencies.
  • Flexible deployment: Supports traditional binaries, Docker, and Nanos Unikernels for minimal VM deployment.

Related

  • Project
    prompt-security/clawsec

    A collection of security skills and signed advisory intelligence for AI agent runtimes to verify artifacts, detect configuration drift, and gate risky installations.

  • Project
    prehisle/relay-pulse

    A service quality monitoring tool for LLM APIs that verifies actual token generation to prevent "fake" uptime reports from proxy services.

  • Project
    erha19/ping-island

    A macOS menu bar app that monitors AI coding sessions, allowing users to approve tools and answer agent prompts directly from the notch surface.

  • Project
    xalgorix/xalgorix

    An autonomous AI penetration testing platform that uses an LLM agent and an independent verifier to discover and prove vulnerabilities with concrete evidence.