AI writing assistant.

The AI Writing Assistant helps Partners create and improve text content in the dashboard. When enabled, an AI button appears in supported text fields, offering quick transformations like improving grammar, adjusting tone, or generating content.

What it does

Supported providers

The system uses an OpenAI-compatible API format, supporting multiple providers:

All providers use the same configuration variables, so you can switch between them.

Configuration

The assistant reads four environment variables:

OPENAI_ORGANIZATION is an optional fifth variable for OpenAI accounts that belong to an organization.

Enable AI features by adding the following to your .env file:

OpenAI

OPENAI_ENABLED=true
OPENAI_API_KEY="your-api-key"
OPENAI_MODEL="gpt-3.5-turbo"
OPENAI_BASE_URL="https://api.openai.com/v1"

DeepSeek

OPENAI_ENABLED=true
OPENAI_API_KEY="your-deepseek-key"
OPENAI_MODEL="deepseek-chat"
OPENAI_BASE_URL="https://api.deepseek.com"

LM Studio (local)

OPENAI_ENABLED=true
OPENAI_API_KEY="lm-studio"
OPENAI_BASE_URL="http://localhost:1234/v1"

After updating your .env file, clear the config cache:

php artisan config:clear

How requests are sent

The assistant calls the chat completions endpoint on the configured base URL. Every request includes a max_tokens limit (a small per-field value) and a preset temperature, so the configured model must accept those parameters. Models that reject them, such as reasoning models with restricted parameter support, fail with an API error.

OpenAI setup

To use OpenAI's GPT models:

1. Visit platform.openai.com and create an account
2. Navigate to API Keys in your dashboard
3. Click Create new secret key and name it for reference
4. Copy the key now. You won't see it again
5. Add the key to your .env file as OPENAI_API_KEY

Choosing a model

When OPENAI_MODEL is not set, the app uses gpt-4. The sample .env ships with gpt-3.5-turbo. Any model that accepts standard chat completions requests with max_tokens and temperature works, whether from OpenAI or from another OpenAI-compatible service configured via OPENAI_BASE_URL.

For the latest models and pricing, see OpenAI's model documentation.

DeepSeek setup

DeepSeek provides an OpenAI-compatible API with competitive pricing:

1. Create an account at platform.deepseek.com
2. Navigate to your API dashboard
3. Generate an API key
4. Add the DeepSeek configuration to your .env file

DeepSeek models

The app recognizes deepseek-chat and deepseek-reasoner by name and adjusts the request format for them. Other DeepSeek model names are sent as regular chat completions requests.

LM Studio setup

LM Studio lets you run open-source AI models locally, ideal for development, testing, or privacy-focused deployments.

Installation

1. Download LM Studio from lmstudio.ai
2. Install and launch the application
3. Search for and download a model (e.g., Llama 3.2, Mistral, or Qwen)

Starting the local server

1. Open LM Studio
2. Go to the Local Server tab
3. Select your downloaded model
4. Click Start Server
5. The API will be available at http://localhost:1234/v1

Recommended local models

For text writing assistance, consider:

• Llama 3.2 8B: Good balance of speed and quality
• Mistral 7B: Fast and efficient
• Qwen 2.5 7B: Strong multilingual support

Model quality depends on your hardware. Start with smaller models (7B–8B parameters) if you have limited RAM.

Troubleshooting

Connection errors

LM Studio issues

General tips

• Clear cache after any .env changes: php artisan config:clear
• Check logs in storage/logs/laravel.log for error details
• Test the API with a curl command before troubleshooting the app

Related topics

• Languages & translations: Multi-language support
• Email configuration: Another .env configuration
• Partners overview: Partner dashboard features