Model Selection
Where to find it
The model selector is located at the bottom-left of the composer area, below the text input.
Auto mode (default)
Auto is the recommended setting. It automatically routes your message to the most suitable model based on the complexity and type of your question.
Available models
Tesseract supports multiple AI models from various providers. The model selector shows each model with its provider icon. Available models may include options from OpenAI, Google, Anthropic, and others, depending on server configuration.
When to switch models
- Most users: Stick with Auto mode — it handles model selection for you.
- Advanced users: Switch to a specific model if you prefer a particular style of response or need a model known for a specific strength (e.g., coding, creative writing, math).
Response Variants (Retry & Forks)
Retry
Below every assistant response, there is a retry button. Clicking it asks the AI to generate a new response to the same message, potentially with different content, structure, or level of detail.
What are forks?
Each retry creates a fork — a variant of the response. The original response and all retries are preserved as separate forks.
Navigating forks
When a message has multiple forks, you’ll see arrow buttons and a counter (e.g., “2 / 3”) below the response:
- Click the left arrow to view the previous variant.
- Click the right arrow to view the next variant.
- The counter shows which variant you’re viewing and how many exist.
Persistence
All forks are saved with the conversation. You can revisit any conversation and navigate between forks at any time — no variants are lost.