Chat API Endpoint

Enable programmatic interaction with Ottogrid’s AI chat assistant. This API allows users to create and modify tables dynamically by sending structured messages, facilitating automated workflows and AI-driven table configurations.

Endpoint

POST https://api.ottogrid.ai/api/v1/bases/ai-configure-base

Headers

• X-API-Key: Your API key

Request Body

{
  "baseName": "Example Base Name",
  "baseId": "optional_base_id", 
  "messages": [
    {
      "role": "user",
      "content": "I want a company name primary column, then using the company name retrieve the company website, another column to search for the CEO name."
    }
  ]
}

Request Parameters

Each message object contains:

• role: Defines the sender. Can be "user", "assistant", or "tool".

• content: The message text. If AI actions are triggered, responses may contain structured objects such as column configurations.

Response

A successful request returns the base ID and a history of exchanged messages.

Example Response

{
  "baseId": "8tjiVYpT",
  "messages": [
    {
      "role": "assistant",
      "content": [
        {
          "type": "text",
          "text": "I'll help you create a workflow with three columns: a document input, its summary, and publication date."
        },
        {
          "type": "tool-call",
          "toolCallId": "toolu_01JHzwSqU6bvReSmGK1yH6Mi",
          "toolName": "manageColumn",
          "args": {
            "columnOperation": "create"
          }
        }
      ]
    },
    {
      "role": "tool",
      "content": [
        {
          "type": "tool-result",
          "toolCallId": "toolu_01JHzwSqU6bvReSmGK1yH6Mi",
          "toolName": "manageColumn",
          "result": {
            "message": "Column created successfully",
            "column": {
              "id": "scol_A6UYtgm5vL988Hbv",
              "name": "Document",
              "type": "file",
              "instructions": "Upload the document you want to analyze",
              "isAiEnabled": false,
              "isPrimary": true,
              "inputColumnIds": [],
              "tools": [
                {
                  "name": "file-analysis"
                }
              ]
            }
          }
        }
      ]
    }
  ]
}

Response Details

Each message object contains:

• role: "assistant", "user", or "tool".

• content: The message body, which can include textual responses or structured tool calls.

Important Note on API Usage

🔹 Be Explicit with Your Requests: When defining table structures through the API, clarity in your instructions is crucial. Clearly outline:

• Primary Column: Specify which column should be the main entity (e.g., Company Name).

• Dependencies: Explicitly state which columns should reference others.

• Tools to Use: If you have a preference on which tools a column should use like scrape-url, include this in your prompt when describing the column.

📌 Example Instruction for Clarity:

"I want a primary column for Company Name. Create a Company Website column using Web Search which uses the Company Name column as input. Then, using Company Website, create a Pricing Model column using the Scrape URL tool."

🔸 Why This Matters:

• AI will follow your structure exactly as described.

• Reduces the risk of incorrect column dependencies.

• Ensures AI selects the correct tools for data extraction.

Capabilities

This API allows you to dynamically interact with Ottogrid’s AI to create and modify tables. You can:

• Create Tables Dynamically: Generate new tables and assign them custom base names.

• Modify Existing Tables: Add, update, or reconfigure columns in an existing base.

• Create New Columns: Define new columns, their types, and dependencies using structured commands.

• Modify Existing Columns: Update instructions, dependencies, or AI settings for existing columns.

💡 Tip: If you are dynamically creating tables, store the baseId from the response to ensure you can modify the same table in future API calls.

Cost Breakdown

• Creating a new table: 5 credits

• General API interactions (e.g., modifying an existing base): 2 credits

Best Practices & Considerations

• Use baseId when modifying existing tables. If omitted, a new base will be created.

• Extract structured outputs from the tool role.

• Store baseId for ongoing API interactions. This ensures consistency when making modifications over multiple API calls.

• Review tool-call results. If creating a column, verify the column details using the correct path in the response. Reference the provided example output to find the exact key structure.