WritingmateWritingmate Documentation

Key Features

Writingmate brings chat, web search, files, images, and video into one workspace. You can choose the model that fits a task, compare several answers, and save useful work for later.

Choose from hundreds of AI models and tools

Use text, image, and video models from providers such as OpenAI, Anthropic, Google, Meta, xAI, Perplexity, and Black Forest Labs. Open the model selector in the message bar, then use its Text, Images, or Video tab to find a model.

Models are grouped by plan tier. Select Manage Models at the bottom of the model selector to choose which text models appear in your list. The available catalog changes as providers release or retire models, so use the AI Models Directory or Model Settings for the current list.

Compare model responses

Compare up to four text models with the same prompt. Writingmate shows each response in its own column, which makes it easier to compare reasoning, writing style, and completeness with your own task instead of relying only on benchmarks.

To start a comparison:

  1. Select Compare Mode in the message bar.
  2. Open the model selector.
  3. Select two to four text models.
  4. Send one prompt to all selected models.

Work with files

Attach a PDF, document, spreadsheet, text file, or supported image to a chat. You can ask for a summary, extract facts, compare sections, or create a new draft from the file.

Select the + button in the message bar, then choose Upload from computer. If your plan includes a knowledge base, you can also choose Add from knowledge base to reuse a saved file. See Uploading and Managing Files and Using Files for details.

Search the web and use connected tools

Web search gives supported text models current information from online sources. Select the + button in the message bar, then choose Web search before sending your prompt.

Writingmate can also connect a chat to outside tools through MCP. MCP is a standard that lets an AI app use services such as GitHub or Slack when a configured server provides those tools. See Web Search and MCP Integration.

Generate and edit images

Select Images in the sidebar or open the Images tab in the model selector. Describe the image you want, choose an image model, and generate it. Depending on the model, you can also upload reference images or edit part of an existing image.

Writingmate supports image models such as Nano Banana, GPT Image, FLUX, Grok Imagine, and Seedream. Model availability depends on your plan and can change over time. See AI Image Generation.

Generate short videos

Select Video in the sidebar to create a video from a text prompt. Supported models can also animate an uploaded image. The Video workspace lets you choose a model, aspect ratio, duration, resolution, and other options that the selected model supports.

Writingmate includes video models such as Sora, Veo, Seedance, Kling, and PixVerse. Your monthly video allowance depends on your plan. See AI Video Generation and Compare Subscription Plans.

Save prompts, projects, and chats

Use Prompts to save instructions you want to reuse. Type / in the message bar to insert a saved prompt into a chat. Projects keep related chats together, while Search helps you find previous conversations.

The sidebar also lists recent and pinned chats. You can share a chat with a link or export it as Markdown, JSON, or CSV from the chat menu. See Sidebar Navigation and Managing Chat History.

Use pre-built or custom Agents

An Agent is an AI assistant with saved instructions, a selected model, optional knowledge files, and optional tools. Use a pre-built Agent for a common workflow, or create one for a repeated task in your own work.

Open Agents in the sidebar to browse, search, or create Agents. See Overview of Agents and Creating Custom Agents.

Agent Mode in the message bar is a separate feature. Use it when a task needs multi-step research, analysis, or planning instead of a single chat response.

Speak prompts and listen to answers

Select the microphone button in the message bar to record a prompt. Writingmate transcribes the recording into text before you send it. On an AI response, use Play Message to listen to the answer or stop playback.

Voice Input availability depends on your plan and requires microphone permission in your browser.

Use Writingmate through an API

Developers can call supported Writingmate models from existing OpenAI clients and tools. A Developer Key is a password-like token that lets another app use your Writingmate account. Add the Developer Key to your client, then set the OpenAI-compatible base URL to https://writingmate.ai/api/openai/v1.

See the OpenAI-Compatible API for supported endpoints and examples.