How should I configure each of the model settings for WriteMapper's Ask AI feature?

A short help guide for how to configure each of the Ask AI model settings.

Name

Sets the display name for the model to remind you of its purpose or intended use, especially useful if you have multiple models saved. Has no impact on actual Ask AI model usage.

Initial System Prompt

Sets the system instruction which the model will use to form its output in your use of Ask AI. See these guides from OpenAI, Google, and Anthropic for more information on how to set effective prompts.

Model Code

Enter the model code you wish to use from your AI provider, for example: "gpt-5", "gemini-2.5-pro", or "claude-opus-4-1". For a full list of model codes, see the respective OpenAI, Gemini, and Claude reference pages.

Thinking / Reasoning

Maps to the Reasoning Effort setting for supported OpenAI models, Thinking for supported Gemini models, and Extended Thinking for supported Claude models. For Gemini and Claude models, "minimal", "low", "medium", and "high" map to a 0-token (disabled), 1,024-token, 8,192-token, and 24,576-token thinking budget respectively (note that reasoning cannot be turned off for Gemini 2.5 Pro models). Configure this to "unset" if using a non-thinking/non-reasoning model, to exclude this setting during usage.

Verbosity

A setting to constrain the verbosity of the model's responses (for OpenAI models only). Configure this to "unset" if using Gemini or Claude, which excludes the verbosity setting during usage.