Job Creation Agent

Using Agents and the MCP Protocol to Create Jobs

We support the Job creation in the MCP Server, allowing you to easily create data collection Jobs using natural language.

📘

Call tools/list to ensure you have the most up-to-date specification

We recommend calling the method tools/list directly on our MCP Server to receive the most up-to-date specification of the Job creation Tool.

Prompt Engineering and Prompt Requirements

Your client, when connected to our MCP Server, give you the ability to create data collection Jobs using natural language. For example:

"Search for dogs and cats on Twitter in the last 7 days"

Within Datastreamer, the natural language prompt will be dissected into parts and used to form data-source-ready jobs for each data source.

Prompts can be complex, and can handle additional specifications like:

• AND/OR operators
• Grouping
• Language
• NOT operator
• Exact phrases

"Create Job" Tool Specification

Using your prompt, your MCP Client will be able to call the create_job tool with the correct parameters and it will create the Job in the specified pipeline.

Below is the tool specification:

{
    "result": {
        "tools": [
            {
                "name": "create_jobs",
                "description": "Create a new job in a component step in an existing Datastreamer pipeline.",
                "inputSchema": {
                    "type": "object",
                    "properties": {
                        "pipelineId": {
                            "description": "The pipeline ID (e.g. d7eadec2)",
                            "type": "string"
                        },
                        "stepId": {
                            "description": "The step ID (e.g. mfootxdv)",
                            "type": "string"
                        },
                        "text": {
                            "description": "The user prompt describing how the new job to be created, including search terms, social networks, time period, etc. (e.g. 'Create a job that searches for cats and dogs on Twitter from last week')",
                            "type": "string"
                        }
                    },
                    "required": [
                        "pipelineId",
                        "stepId",
                        "text"
                    ]
                }
            }
        ]
    },
    "id": 1,
    "jsonrpc": "2.0"
}

🚧

Early Access Additional Requirements

As the Job Creation Agent is in early-access, two additional fields are currently required:

• Pipeline ID (Temporary)
• Step ID (Temporary)

Meaning the above example would be: "Searches for dogs and cats on Twitter in the last 7 days, in the pipeline_id 30601121 and step_id 0erc2jx4"

Supported Providers and Sources

Currently supported providers and sources include:

• Socialgist (select sources)
• Vetric (select sources)
• WebSightLine
• Data365 (select sources)

We are continuing to rollout more supported element, as our team is able to verify compatibility tests with each source and provider.