Getting Started with DIMARC
Introduction to DIMARC
DIMARC is a French artificial intelligence platform that offers specialized services for processing, analyzing, and generating content. Designed for businesses and developers, DIMARC provides turnkey solutions to integrate advanced AI capabilities into your applications and business processes.
The platform stands out for:
- French and European AI models complying with GDPR standards
- An intuitive interface for creating and managing your AI agents
- Simple APIs to integrate into your existing applications
- Specialized solutions for specific use cases
Account Creation and Configuration
Create a DIMARC Account
- Go to the DIMARC portal
- Click on the “Sign up” or “Create an account” button
- Fill out the registration form with your information
- Accept the terms of use
- Click on “Create my account”
Account Validation
Once your registration is complete:
- Check your inbox for the confirmation email
- Click on the confirmation link in the email
- Your account is now activated
First Login
- Log in to your dashboard
- Take time to explore the interface and different sections
- Complete your user profile and organization information if necessary
Understanding DIMARC Agents
What is a DIMARC Agent?
A DIMARC agent is an AI instance configured to address a specific need. Each agent has:
- A precise objective
- Its own knowledge (documents, databases)
- Customization parameters
- A dedicated API interface
Types of Available Agents
DIMARC offers several specialized AI services:
Service | Description | Use Cases |
---|---|---|
Documentalist | Analyzes and answers questions about your documents | Document chatbot, automated FAQ, customer assistance |
Extractor | Extracts structured information from unstructured documents | Invoice processing, contract analysis, data extraction |
Writer | Coming soon. Generates quality text content in seconds | Product description creation, article writing, text generation |
How to Choose the Right Service
To select the appropriate service, ask yourself these questions:
- Do you want to analyze existing documents? → Documentalist
- Do you need to extract specific data from documents? → Extractor
- Do you want to generate new content? → Writer
Creating Your First Agent
Agent Creation Steps
- In your dashboard, click on “Create an agent”
- Select the desired agent type (Documentalist, Extractor, or Writer)
- Give your agent a meaningful name
- Configure the specific parameters for the type of agent.
Advanced Configuration
Depending on the type of agent chosen, you will be able to:
For the Documentalist:
- Import your documents (PDF, DOCX, TXT)
- Define external knowledge sources
- Set the response style and response constraints
For the Extractor:
- Define fields to extract
- Create custom extraction templates
For the Writer:
- Define content templates
- Set tone and voice
- Configure length and format constraints
Visibility and Sharing
You can set 3 visibility levels:
Level | Description |
---|---|
Private | Accessible only by you and administrators |
Organization | Accessible by all members of your organization |
Public | Accessible by everyone via a dedicated URL |
Testing and Validation
Before making your agent available:
- Submit several queries to verify the responses
- Adjust parameters if necessary
- Validate the final version of your agent
Quota Management
Each DIMARC plan includes a usage quota:
- Track your consumption in the “Usage” section of your organization’s dashboard
Basic Integration
Technical Prerequisites
To integrate the DIMARC API, you will need:
- A valid API key
- A development environment capable of making HTTP requests
- Basic programming knowledge
Integration Examples
# Example with the Documentalist servicecurl -X POST https://api.dimarc.ai/v2/documentaliste/<agent_id> \ -H "x-api-key": <your_api_key> \ -H "Content-Type: application/json" \ -d '{ "query": "What does Dimarc do?", "instructions": ["Add 🌈 in your responses"], "historic": [] }'
// Example with the Documentalist serviceasync function queryDocumentalist(question) {const response = await fetch('https://api.dimarc.ai/v2/documentaliste/<agent_id>', {method: 'POST',headers: {'x-api-key': '<your_api_key>','Content-Type': 'application/json'},body: JSON.stringify({query: "What does Dimarc do?",instructions: ["Add 🌈 in your responses"],historic: []})});
return await response.json();}
// UsagequeryDocumentalist('What is the refund procedure?').then(result => console.log(result)).catch(error => console.error(error));
# Example with the Documentalist serviceimport requests
def query_documentalist(question):url = "https://api.dimarc.ai/v2/documentaliste/<agent_id>"headers = { "x-api-key": "<your_api_key>", "Content-Type": "application/json"}data = { "query": "What does Dimarc do?", "instructions": ["Add 🌈 in your responses"], "historic": []}
response = requests.post(url, headers=headers, json=data)return response.json()
# Usageresult = query_documentalist("What is the refund procedure?")print(result)
Practical Use Cases
Documentalist for a FAQ
Use the Documentalist to create an intelligent FAQ on your website:
- Create a Documentalist agent
- Import your support documents, terms of use, etc.
- Integrate the API into your website with a chat interface
- Allow your users to ask questions in natural language
Extractor for Document Processing
Automate information extraction from invoices:
- Create an Extractor agent
- Configure the fields to extract (date, amount, invoice number, etc.)
- Submit your invoices via the API
- Retrieve structured data in JSON to integrate with your accounting system
Writer for Content Generation
Automatically generate product descriptions:
- Create a Writer agent
- Define a template for your descriptions
- Provide the technical characteristics of the product
- Get a professional marketing description
Best Practices and Tips
Performance Optimization
- Be as specific as possible in your queries to get more relevant responses
- Limit the size of imported documents to relevant sections
- Intelligently manage conversation histories for chat applications
Cost Reduction
- Optimize your queries to limit the number of tokens used
- Cache frequent responses
- Regularly monitor your consumption
Improving Response Quality
- Progressively train your agent by providing feedback
- Regularly update your document bases
- Test different phrasings for the same questions
- Use performance metrics to identify areas for improvement
Additional Resources
Technical Documentation
- API Reference: Detailed API documentation
- Specific Guides: Tutorials for each service
- Technical FAQ: Answers to frequently asked questions
Support and Community
- Technical Support: Contact our team
Next Steps
Now that you master the basics of DIMARC, you can:
- Explore advanced features of each service
- Integrate multiple agents to create complex workflows
- Set up analysis and feedback systems
- Participate in our beta program to test new features