This project was built as part of the Crustdata API Chatbot Challenge, where I developed a chatbot named CrustBot that assists users in querying Crustdata's APIs. The chatbot was designed and implemented incrementally through multiple levels of functionality, as outlined below. 🚀

• Implemented a basic chat interface where users can ask questions about Crustdata's APIs.
• The chatbot provides accurate technical answers based on the API documentation.
• Examples of supported queries include:
  • Searching for people by title, company, and location.
  • Understanding API error responses and resolving issues.
• Status: ✅ Completed and deployed.

• Validates API request examples before sharing them with the user.
• Automatically checks for errors in API requests and suggests corrections.
• Supports conversational mode, allowing users to ask follow-up questions within the same thread.
• Status: ✅ Implemented and functional.

• Added the ability to ingest and integrate additional knowledge bases, including:
  • Questions and answers from Crustdata’s support Slack channels.
  • Updated API documentation.
  • Additional resources to improve chatbot accuracy.
• Status: ✅ Integrated.

• Integrated the chatbot into Slack as a Slack bot:
  • Works within specific channels.
  • Accessible to designated users only.
  • Drafts responses for messages in real-time.
• Status: ✅ Fully functional Slack bot deployed.

The complete code for this project is available on my GitHub: Crustdata Chatbot Repository

Question: "How do I search for people given their current title, current company, and location?"

Answer: Use the endpoint api.crustdata.com/screener/person/search. Here’s a sample cURL request:

curl --location 'https://api.crustdata.com/screener/person/search' \
--header 'Content-Type: application/json' \
--header 'Authorization: Token $token' \
--data '{
   "filters": [
       {
           "filter_type": "CURRENT_COMPANY",
           "type": "in",
           "value": [
               "openai.com"
           ]
       },
       {
           "filter_type": "CURRENT_TITLE",
           "type": "in",
           "value": [
               "engineer"
           ]
       },
       {
           "filter_type": "REGION",
           "type": "in",
           "value": [
               "San Francisco, California, United States"
           ]
       }        
   ],
   "page": 1
}'

Question: "I get an error when using the region filter. How should I format my input?"

Answer: Ensure that the region value matches one of the supported options from the list here: Updated Region Values.

For any queries, feel free to reach out via LinkedIn or open an issue on the repository.

Thank you for reviewing my project! 😊