Model keys and credits

The agent needs an AI model to write code. Emolysis gives you two ways to provide one: your own key from the model provider you choose, or a paid plan's model usage. Every plan accepts your own model key (on Enterprise, an organization key), and the Free plan works only that way. Prices and plan contents are on the pricing page.

At a glance #

Bring your own key #

Create a key with the model provider you choose and add it to Emolysis once, in the app or with the command line. The key is kept in your machine's credential store, never in a project file. The agent sends its requests to that provider under your key; the provider bills you directly, and its terms apply to your requests. Emolysis charges nothing for model usage with your own key, and those requests do not count against your plan's limits. You manage the key, its spending limits and its rotation yourself.

Bring your own key is how the Free plan works. Paid plans accept it too, so you can move to a paid plan for connectors, sharing and checks and keep your own key. On Enterprise, admins can set one organization key for every member. See bring your own key in the glossary.

Included usage and usage credits #

If you would rather not manage a key, Pro, Max and Team include model usage, so you do not need a key. Included usage is measured by the work the agent does, not by a fixed message count. Pro sets the baseline, Max gives 5x or 20x Pro, a Team standard seat gives more than Pro, and a premium seat gives 5x a standard seat. Limits apply to each person. The exact allowances are published with the paid plans.

Past a limit, usage credits keep the agent working at standard metered rates, until you turn them off or reach a spend limit you set. Usage credits are part of Pro, Max and Team.

Enterprise has no included allowance. Model usage is billed at standard metered rates from the first request, within spend limits that admins set for each user and the organization, unless the organization uses its own model key.

Requests that use a paid plan's model usage pass through Emolysis's metering service on their way to the model provider. That service is designed not to store or log the content of prompts or responses; it records token counts so usage can be billed against your plan.

When you reach a limit #

You have four options: wait for the limit to reset, switch the agent to your own key, move to a plan with more included usage, or turn on usage credits. When no included usage or credits are left, model requests stop with an insufficient_credits error. Nothing in your project is lost. Work that does not need the model, such as golden tests on DuckDB, keeps working. See Error codes.

What the provider sees #

In both cases the model provider receives the agent's prompts. A prompt can include table and column names, schemas, the template being filled and sample values from your data. See Data handling for the full list of what leaves your machine, and when.

Choosing a provider #

Emolysis works with several model providers. We do not recommend one on this site. Choose based on your company's agreements with providers and their data terms. If your company restricts which providers its data may reach, check that before you connect a platform, because sample values can appear in prompts.

Changing keys #

You can replace or remove your key at any time. Removing it stops model requests until you add another key or, on a paid plan, switch to the plan's model usage.

If your editor brings its own model #

An editor or coding agent that supports MCP uses its own model provider and its own key to talk to you. Emolysis's tools return their results to it, and it sends them on under its own terms. See The MCP server.

Status #

Plans and model usage are not on sale yet. Until paid plans open, nothing can be bought.