Supported Models
CraftCore supports three cloud providers plus custom OpenAI-compatible endpoints. Models are auto-fetched from each provider's API and cached for 24 hours. Fallback lists are used when offline.
OpenAI
Models are fetched live from the OpenAI API. Common models:
| Model | Best For | Notes |
|---|---|---|
gpt-4.1-mini | Translation, simple generation | Fast, cost-effective |
gpt-4.1 | Complex dialogue writing | Higher quality |
gpt-4.1-nano | High-volume batch tasks | Fastest, cheapest |
o4-mini | Reasoning-heavy tasks | Reasoning model |
Anthropic
| Model | Best For | Notes |
|---|---|---|
claude-haiku-4-5 | Translation, simple generation | Fast, cost-effective |
claude-sonnet-4-5 | Complex dialogue, nuanced writing | Best balance |
claude-opus-4 | Complex creative writing | Highest quality |
Google (Gemini)
Models are filtered to the latest generation. Only base models (pro, flash, flash-lite) are shown.
| Model | Best For | Notes |
|---|---|---|
gemini-2.5-flash | Balanced performance | Fast with reasoning |
gemini-2.5-pro | Complex tasks | Highest quality, 1M context |
gemini-2.0-flash | Simple generation | Ultra-fast |
gemini-2.0-flash-lite | Batch tasks | Most efficient |
Custom Endpoints
Any OpenAI-compatible chat completions API works as a custom provider. This includes:
- Self-hosted models -- Ollama, LM Studio, vLLM, llama.cpp
- Alternative providers -- Together AI, Groq, Fireworks, Deepseek
- Corporate gateways -- Azure OpenAI, AWS Bedrock proxies
The custom client uses the same OpenAIClient with a custom base URL.
Model Selection Tips
For Translation
Use cheap, fast models: gpt-4.1-mini, claude-haiku-4-5, gemini-2.0-flash. Translation quality is consistent across tiers.
For Dialogue Writing
Use mid-tier or higher: gpt-4.1, claude-sonnet-4-5, gemini-2.5-pro. Better understanding of character voice and narrative structure.
Temperature Guidelines
| Temperature | Use Case |
|---|---|
| 0.0 -- 0.3 | Translation (consistent, faithful) |
| 0.3 -- 0.6 | Editing, modification |
| 0.6 -- 0.9 | Creative generation |
Usage Tracking
ApiUsageTracker automatically tracks tokens and estimated costs across all API calls. View usage in the CraftCore Hub Dashboard tab.