IPC API Reference

February 11, 2026 · View on GitHub

All communication between the renderer process and main process is done via window.api.*, returning Promise<IpcResponse>.

Response Format

All API calls return a unified response format:

interface IpcResponse<T = any> {
  success: boolean;
  data?: T;
  error?: string;
}

Providers

Manage LLM service providers (OpenAI, Anthropic, etc.).

Method	Description
`window.api.provider.getAll()`	Get all providers
`window.api.provider.getById(id)`	Get provider by ID
`window.api.provider.create(data)`	Create a new provider
`window.api.provider.update(id, data)`	Update an existing provider
`window.api.provider.delete(id)`	Delete a provider
`window.api.provider.updateStatus(id, status)`	Update provider status

Models

Manage LLM models associated with providers.

Method	Description
`window.api.model.getAll()`	Get all models (grouped by provider)
`window.api.model.getByProvider(providerId)`	Get models by provider ID
`window.api.model.create(data)`	Create a new model
`window.api.model.delete(id, provider)`	Delete a model

Tasks

Manage PDF/image conversion tasks.

Method	Description
`window.api.task.create(tasks[])`	Create tasks (batch)
`window.api.task.getAll({page, pageSize})`	Get tasks (paginated)
`window.api.task.getById(id)`	Get task by ID
`window.api.task.update(id, data)`	Update a task
`window.api.task.delete(id)`	Delete a task
`window.api.task.hasRunningTasks()`	Check if there are running tasks

Task Details

Manage individual page conversion details within a task.

Method	Description
`window.api.taskDetail.getByPage(taskId, page)`	Get task detail with image for a page
`window.api.taskDetail.getAllByTask(taskId)`	Get all task details for a task
`window.api.taskDetail.retry(pageId)`	Retry a single failed page
`window.api.taskDetail.retryFailed(taskId)`	Retry all failed pages in a task

Files

Handle file operations (upload, download, dialog).

Method	Description
`window.api.file.selectDialog()`	Open file selection dialog
`window.api.file.upload(taskId, filePath)`	Upload file
`window.api.file.uploadFileContent(taskId, fileName, fileBuffer)`	Upload file content as ArrayBuffer
`window.api.file.getImagePath(taskId, page)`	Get image path for a page
`window.api.file.downloadMarkdown(taskId)`	Download merged markdown file

Completion

LLM completion and testing operations.

Method	Description
`window.api.completion.markImagedown(providerId, modelId, url)`	Convert image to markdown
`window.api.completion.testConnection(providerId, modelId)`	Test model connection

Shell

System shell operations.

Method	Description
`window.api.shell.openExternal(url)`	Open URL in default browser

Window

Window control operations.

Method	Description
`window.api.window.minimize()`	Minimize the window
`window.api.window.maximize()`	Maximize/restore the window
`window.api.window.close()`	Close the window

Events

Subscribe to real-time events from the main process.

Method	Description
`window.api.events.onTaskEvent(callback)`	Subscribe to task events (returns unsubscribe function)
`window.api.events.onTaskDetailEvent(callback)`	Subscribe to task detail events (returns unsubscribe function)

Usage Example

// Subscribe to task events
const unsubscribe = window.api.events.onTaskEvent((event) => {
  console.log('Task event:', event);
});

// Unsubscribe when done
unsubscribe();

App

Application information.

Method	Description
`window.api.app.getVersion()`	Get app version from package.json

Platform

Platform information (synchronous property).

Property	Description
`window.api.platform`	Get current platform (`'win32'`, `'darwin'`, `'linux'`)