What OpenLLM proposes
OpenLLM, presented by co-founder Jonathan Archer Hunter, introduces a unified API key intended to let users combine subscriptions, API keys and local models across coding agents and the ChatGPT app on macOS. The project is described as enabling remote coding sessions that start on a home machine from a laptop, the web app or the Mac app when the target machine is online.
The team says the platform supports a fallback chain so models are tried in order when one hits a limit, and provides a single shared memory so decisions or preferences saved once can be recalled by any connected app. It also reports which model answered a request and how many tokens were consumed.
Supported integrations and free options
According to the announcement, the same OpenLLM key will work in apps that accept OpenAI or Anthropic keys. The launch lists compatibility with Claude Code, Codex, Kimi Code, Grok, Cursor and Muse Code, as well as API keys, custom API keys, local models and a Brave Search key. The team notes free inference options, including DeepSeek and Qwen, so an initial request can run before user accounts are connected.
OpenLLM also mentioned free context.dev credits to allow agents to read the web, and said an MCP marketplace is planned.
Limits, pricing and team features
OpenLLM is free to start. The announced Pro plan costs $30 per month for up to five devices and includes a three-day trial. A promotional code (PHUNT) was offered for 50% off Pro for six months to the first 100 users. The release notes that the service runs on macOS and Linux.
The company is preparing team capabilities that will let administrators set token limits per person, observe which apps team members use, distribute the same tools and MCPs and share a single memory across accounts. The team emphasised that OpenLLM does not make any subscription unlimited, is not intended for account sharing, and that vendors’ limits and terms still apply to users’ own accounts on their own machines.
Documentation and sign-up details are available from OpenLLM; the co-founder said the team has answered common questions such as whether using the service could affect account status.
Original source: Product Hunt