Meet ModelMesh. More models, one workspace.
How the catalog, API gateway, and dashboard fit together — and where to start.
Start with a model, not another integration
ModelMesh brings available models into one catalog with an exact API identifier, input and output prices, and published capabilities. Your application connects to the gateway, while your workspace keeps keys, billing, and request history together. Pick a model for the job you need to do, then keep the same integration when you want to try another.
Keep the tools you already use
An OpenAI-compatible client can use the ModelMesh base URL and your API key. The Connections page also covers Codex CLI through the Responses API and Claude Code through the Messages API. Choose an available model suited to the tool: coding assistants need tool-calling support, while a simple text application may only need chat completions.
Know what a request costs
Read both prices in the catalog before choosing a model. Long prompts increase input usage; long responses increase output usage. The pricing calculator estimates a request using the selected model's current rates. Your actual requests and recorded costs appear in Usage and Request logs.
Build a small first integration
Create an account, generate a key, and try a short request. Confirm that the response and request log look right before adding streaming, tool calls, or a longer conversation. Keep development and production keys separate so changing one integration does not interrupt the other.