OpenLLM

Under review

Your subscriptions, local models and APIs, one key anywhere.

Sends coding agent requests to your subscriptions, local models, or API keys, with fallback chains that move to the next model when one is rate-limited, so work keeps moving.

Built by Jonathan Archer · Artificial Intelligence · Developer Tools · Paid

Coming soon on Launch Llama Tools

Community listing

Listing status

0upvotes

Community score

Jonathan Archer (Joined Nov 2026)

Founder

Compared against 5 alternatives

Market comparison

Sends coding agent requests to your subscriptions, local models, or API keys, with fallback chains that move to the next model when one is rate-limited, so work keeps moving.

OpenLLM demo video

About

OpenLLM puts all your AI under one API key. You can connect local models, ChatGPT and Claude subscriptions you already pay for, and any API provider. Then use that key in any app that accepts an API key or a custom base URL, including coding agents, code review tools, n8n, an always-on VPS, or production.

Instead of paying per token in every app, you route requests to what you already have and save money. Switch models anytime without changing your setup, and see usage and costs in one place.

OpenLLM Teams and Sign in with OpenLLM are coming soon. Teams lets you share inference with your team, friends or family, set usage caps, and share MCP tools and memories.

Pro is $30/mo and comes with a 3-day free trial.

Competitors: litellm, omniroute, llmgateway

Follow OpenLLM: X (Twitter) · LinkedIn

Key Features

Routes Claude Code, Codex, and OpenCode through one local routing layer
Falls back to the next model in your chain when one is rate-limited
Runs local models from Ollama and LM Studio alongside cloud provider APIs
Includes inference so your first request works with no keys connected
Encrypts provider keys in your browser under a recovery phrase you hold
Shares memory and marketplace integrations across every agent you run

Use Cases

Keep a coding agent running mid-task by switching to the next model in your fallback chain when one is rate-limited
Run Claude Code or Codex against a local Ollama or LM Studio model instead of a paid cloud API

From the founder

Who it's for

“For teams and people looking to lower their AI inference expenses. With OpenLLM you can bring your tokens where you want.”

Why I built it

“We built out the product due to our personal need. We started working and hitting limits with our AI subscriptions. We needed a tool that could help us use multiple subscriptions together wherever we wanted. Since we couldn't find any viable options, we built it out ourselves.”

— Jonathan Archer, founder of OpenLLM

Who it is for

Developers who use Claude Code, Codex, or OpenCode alongside several subscriptions, local models, and API keys
Engineers who want one routing layer across multiple devices and model providers