Built by one person, for people who actually work.

The Story

I built Arcochat because I was tired of paying markups on tokens, switching between six different LLM browser tabs, and losing project context every time I started a new chat. I wanted a clean, fast, local-first workspace that stays out of my way.

Arcochat is designed to be completely transparent. By connecting your own API keys, you only pay for exactly what you consume, with zero markups or artificial restrictions from us.

Current State

Arcochat is in active early access. A focused group of developers, founders, and consultants are using the desktop client during this phase, and every shipping decision is driven directly by their feedback. If you join now, you're not a user you're a co-builder. Your workflow shapes what gets built next.

Active Roadmap

We are continuously shipping updates based on the early access queue. Here are the core modules currently in development:

• In Progress
  Local LLM Integrations (Ollama): Run local Llama 3 or Mistral models natively inside sandbox projects.
• Up Next
  Context Routing: Intelligent local embeddings router that automatically pre-selects the most relevant files based on user prompts.
• Planned
  Shared Team Templates: Export and sync project configurations and custom prompt templates across team seats.

Have feedback or questions?

I respond directly to all queries and early access feature suggestions.