Integrations
Designed for any vendor. Owned by none.

OmniConflux is designed to work with any vendor: your subscription's own tool, your API key or local runtimes, any command-line agent, and open protocols like MCP and OAuth 2.1.
Open by design
OmniConflux is designed to integrate with any vendor and never pigeonhole you: models come three ways, any command-line agent is designed to run, and models and harnesses are designed to be swappable without moving your work. Sign-in is designed around OAuth 2.1 and OpenID Connect with PKCE, exact redirect matching, header-bound tokens and refresh rotation. OmniConflux is a desktop coding app for Omarchy, macOS on Apple silicon and Windows, version 0.3.0, in closed beta.

Models through your subscription
The vendor's own installed tool is designed to run on your machine, under your login, as the vendor allows.
- Claude Code
- Codex CLI
- Gemini CLI

Models through your API key
Designed so your key goes direct to the vendor, and to work with any provider you hold a key for, including endpoints compatible with the OpenAI API.
- Anthropic
- OpenAI
- Mistral
- DeepSeek
- xAI

Local models
Open-weight models on your own hardware. The work is designed to stay on the machine.
- Ollama
- llama.cpp
Agent harnesses
Designed for process-level control of any command-line agent, with deeper control, including state, handoff and approvals, for agents that speak MCP or ACP.
Protocols
Open standards, designed so you can check what is happening and swap models and harnesses.
- Claude Code
- Codex CLI
- Gemini CLI
- Aider
- OpenCode
- Any command-line agent

- MCP
- A2A
- ACP
- OAuth 2.1
- OpenID Connect
- PKCE
What we do not claim
By design, we do not run many cloud agents off one personal seat, pool personal accounts across servers, or pass your seat login through a hosted backend. We do not promise zero setup for every agent, or perfect memory shared between vendors. A managed fleet, if you want one, is planned, not live, and will run on your API keys or enterprise tokens.
See how it fits
Read the fleet architecture guide, or see how social posting works on the social page.
Questions
Does OmniConflux work with my AI vendor?
It is designed to. Bring the vendor's own command-line tool on your machine, your API key, or a local model. If an agent runs from the command line, the fleet is designed to run it.
Can I run many cloud agents on one personal subscription?
No. A personal subscription is designed to run through the vendor's own tool on your machine, as the vendor allows. By design, we do not pool personal accounts or pass seat logins through a hosted backend. For scale, use API keys or enterprise tokens.
Do agents from different vendors share memory?
By design, they share work through git worktrees and files. We do not claim lossless context transfer between vendors.
Does a vendor endorse OmniConflux by being listed?
No. A listing describes how OmniConflux is designed to work with that product. It does not imply a partnership or endorsement.
All product names are trademarks of their owners.
