Assistant API integration
The Assistant API provides a natural language AI conversational interface with direct access to MAPEGY’s innovation data (including organization data), eliminating the need to map specific use cases to raw API functions manually.
The API returns JSON. The answer of the assistant is formatted as HTML.
How to Integrate
You can integrate MAPEGY's intelligence directly into your internal applications and workflows in three main ways:
- Directly as a conversational chat interface. Build a dedicated chat window inside your company portal, intranet, or messaging apps.
- Indirectly via another LLM: If you already use an enterprise AI assistant (e.g., a custom internal ChatGPT or Claude agent), you can plug the Assistant API in as a specialized tool or via our MCP server. When a user asks your internal AI a market intelligence question, it automatically calls our Assistant API in the background to retrieve MAPEGY data and incorporates it into its final answer.
- Via automated, predefined prompts: Trigger automated intelligence workflows without requiring a user to manually type a prompt every time. Your application backend can automatically inject dynamic variables into a prompt template (e.g., "Summarize patent filings for {Company_Name} over the last 12 months") and call
/Assistantto populate executive dashboards, weekly email digests, etc.
Getting Started
Before integrating the Assistant API, you must authenticate and generate a token. You can test the API using the API Demo page, where you can also download a JavaScript demo app to test it directly in your browser.
Additionally, you can use the MCP server to test MAPEGY APIs.
Make Requests
To interact with the API, send HTTP GET or POST requests to the API endpoint using the default content type. Every request requires a valid authentication token.
- API Endpoint:
https://api.mapegy.com - Route:
/Assistant - Content-Type:
application/x-www-form-urlencoded - Response Format: The API returns JSON. The answer of the assistant is formatted as HTML.
Required Parameters:
token: Your API access token.message: A question or prompt in natural language (e.g., "Find patents related to solid-state batteries."). For optimal results, add context when available: describe the underlying business problem or strategic opportunity, mention the organization name, department, or job title of the recipient, and specify focus criteria like timeframes or regions.
Optional Parameters:
thread_id: The thread ID of an existing conversation to continue a multi-turn chat. Omit this parameter to start a brand-new thread.
Conversation threads are automatically deleted 30 days after the last activity.
Example response
{
"message": "<p>Here is an analysis of patents related to solid-state batteries...</p>",
"used_conversation_length_percent": 9.5785,
"thread_id": 429
}