Editing
Editing features are enabled in the editor app. With them, you can:
Write text from scratch. For example, when it is hard to come up with an article structure.
Transcribe speech to text. This can be a call recording or simple voice dictation.
Format existing text. For cases when it needs to be shortened, made more formal, or better structured.
Requirements
OpenAI API-compatible inference server:
Cloud services: OpenAI, DeepSeek.
Local or open-source models: LM Studio, Llama.cpp, vLLM, or another server that supports the OpenAI API.
How it works
The Gramax app sends a request (text or audio) to the
gramax/editor-aiAI server running in Docker.The
gramax/editor-aiAI server receives the request and transforms it into a standard request for the selected model.The model performs the task — generates text, transcribes audio, or edits content.
The result is returned through the
gramax/editor-aiAI server back to the Gramax app.
Setup
Deploy Docker Compose with the rewriting server as described here — https://hub.docker.com/r/gramax/editor-ai.
Minimal docker-compose.yaml:
version: "3.8" services: editor-ai: image: gramax/editor-ai environment: - CHAT__TYPE=openai - CHAT__MODEL=gpt-4o - CHAT__APIKEY=<your API key> - AUTH__ADMIN__TOKEN=<token for Gramax> ports: - 3005:3005 volumes: - ./config.yaml:/app/config/config.yaml
Connect AI
In space settings, click AI server.
Fill in the server details.
URL of the server where https://hub.docker.com/r/gramax/editor-ai is deployed.
AI server token — set in Docker Compose using the
AUTH__ADMIN__TOKENvariable.
Settings are applied to a specific space on a specific device.
Set a prompt
You can prepare a prompt that rewrites text. For example, define documentation Tone of Voice and general writing requirements in it. To avoid pasting this prompt every time, you can save it directly in the catalog.
In catalog settings, click AI prompts.

Set the prompt and publish it.
The prompt is saved in the catalog and becomes available to all editors.