AI, Screph Code and automation
These parts of the product use project data differently and run separately. Saving the project does not itself issue a LLM request, change the code, or perform any input actions.
Automation Runtime Implemented
Executes a separately prepared GUI script, searches for visual and text states and uses the selected input method. The script runs with the rights of the running program and requires monitoring.
Preparation and safe launch →Agent Experimentally
Offers explanations and supported changes in the context of the current task. The feature is disabled by default; changing actions are shown before being applied and may require confirmation.
Context, permissions and history →Screph Code Implemented
Gets the selected project data and working code folder. Builder can modify files in this folder. Before the request, a backup copy or version control system is needed, after the response, a check of created, modified and deleted files is needed.
Running the code is a separate command and does not start automatically after the Builder response.
Working with Screph Code →Local and cloud models
Локальные профили Ollama и LM Studio обрабатывают разрешённый контекст в локальной службе. Облачные профили передают контекст выбранному провайдеру. Профили назначаются отдельно для Screph Code, Agent и других функций.
Connections and data transfer →Pre-flight checklist
- The project and working files have been saved.
- Checked what data and images will be transferred.
- An available model and the required dependencies have been selected.
- For the changing action, a result check and a rollback method are prepared.