bigdata_create_monitor

Overview

The bigdata_create_monitor tool creates a search monitor in your Bigdata.com account from a plain-language intent. Every other field is optional. Anything you set is pinned, so it is used as given instead of being drafted from the intent. Create always saves the monitor inactive. Preview it, read its runs, then turn it on with bigdata_edit_monitor. The returned monitor id is the handle for every follow-up: fetch it with bigdata_get_monitor, preview it with bigdata_simulate_monitor, read its results with bigdata_get_monitor_runs, change or activate it with bigdata_edit_monitor, or remove it with bigdata_delete_monitor.

When to Use

Create Monitor is ideal for:

Tracking events: Watching for regulatory actions, rating changes, management changes, or any other recurring event
Watching a set of companies: Running the same question on a schedule across a list of entities
Extracting structured fields: Turning each match into a fixed set of fields, such as company, amount, and stage

Entities must be RavenPack entity IDs, not company names or tickers. Resolve them with find_securities first.

How It Works

Create the monitor with an intent and any fields you want to pin. It is saved inactive
Show the returned configuration to the user
Preview it with a flash run of bigdata_simulate_monitor
Read the results with bigdata_get_monitor_runs
Activate it with bigdata_edit_monitor by setting status to active. Add email delivery only if the user asked for it

Parameters

Parameter	Type	Required	Description
`intent`	string	Yes	What the monitor should track, in plain language.
`name`	string	No	Display name.
`search_queries`	string[]	No	Raw fast-mode `bigdata_search` JSON strings. See Search queries.
`schedule`	object	No	Exactly one of `frequency` or `cron`. See Schedule.
`extraction_instructions`	string	No	Extra instructions for what to keep or drop.
`entities`	string[]	No	RavenPack entity IDs. See Entities.

{
  "entities": ["IWGWUR"],
  "intent": "Track regulatory scrutiny of Anthropic PBC: any government, regulator, legislator or competition authority action aimed at Anthropic. Capture new or escalating investigations, probes, inquiries and information requests (e.g. FTC, DOJ, SEC, CMA, European Commission, EU AI Office, national data protection authorities), antitrust or merger reviews of Anthropic deals and investments, enforcement actions, fines, consent orders, regulator-filed lawsuits, government designations or restrictions (e.g. procurement bans, supply-chain risk designations, export controls), congressional or parliamentary hearings and letters targeting Anthropic, and compliance obligations or findings under AI laws such as the EU AI Act. For each event extract: the regulator or government body, jurisdiction, type of action, status/stage, the issue at stake, potential penalties or consequences, and the date.",
  "name": "Anthropic – Regulatory Watch",
  "schedule": {
    "cron": "0 8 * * 1-5",
    "timezone": "Europe/Madrid"
  }
}

Here only entities, name, and schedule are pinned. The search queries and the structured output schema are drafted from the intent. See the response.

Important Notes

The monitor is always created inactive
Pinned fields are used as given and are not drafted from the intent
structured_output and delivery cannot be set at creation. The schema is drafted from the intent. Change it, or add delivery, with bigdata_edit_monitor
If the monitor cannot be generated, the tool returns the rejection text from the API
A smart-mode or malformed search query is rejected

Shared Shapes

These shapes are used by bigdata_create_monitor and bigdata_edit_monitor.

Schedule

Send exactly one of frequency or cron. timezone is only allowed with cron and defaults to UTC.

{"frequency": "6h"}

{"cron": "30 9 * * 1-5", "timezone": "America/New_York"}

frequency is a count plus h, d, or w (for example 4h, 2d, 1w), at most 31 days
cron has 5 fields. The minute must be a single value, so runs are at least an hour apart
A cron can set the day of month or the day of week, not both, and must run at least once every 31 days
Responses rebuild the unit from the stored hours, so 48h can come back as 2d

Search queries

A list of raw JSON strings. Each string is one fast-mode bigdata_search payload. Smart mode is rejected.

["{\"search_mode\": \"fast\", \"query\": {\"text\": \"A regulator has opened an investigation into Anthropic.\"}}"]

Before storing, the MCP server:

Drops filters.timestamp, max_chunks, and reranking. The reranker is stored disabled
Merges payloads that share filters into one multi-text search
Wraps each search as {"query": {...}}

The example above is stored as:

[
  {
    "query": {
      "text": "A regulator has opened an investigation into Anthropic.",
      "ranking_params": {"reranker": {"enabled": false}}
    }
  }
]

Entities

RavenPack entity IDs from find_securities. When set, the list must contain at least one ID, up to the maximum watchlist size. Duplicate IDs are collapsed. The IDs are stored as one watchlist group:

["IWGWUR"]

{"watchlists": [{"entities": ["IWGWUR"]}]}

Structured output

A flat map of field name to {description, field_type, values, read_only}. Only field_type is always required.

field_type is one of entity_reference, number, enum, ordered_enum, string, boolean, datetime, list, range, currency, reference_period, or text
values is required for enum and ordered_enum
The map ends with the read-only system fields announcement_date, effective_date, and summary

{
  "regulator": {
    "description": "The proper name of the government body, regulator, legislature, committee or competition authority taking the action.",
    "field_type": "entity_reference"
  },
  "action_type": {
    "description": "The kind of regulatory action taken against Anthropic.",
    "field_type": "enum",
    "values": ["investigation", "merger_review", "enforcement", "designation_restriction", "legislative_scrutiny", "compliance_obligation"]
  },
  "status": {
    "description": "How far the action has progressed: opened when it is announced, filed or begun, concluded when it ends with a decision, outcome or taking effect; withdrawn if dropped without an outcome.",
    "field_type": "ordered_enum",
    "values": ["opened", "concluded", "withdrawn"]
  }
}

Delivery

The only accepted value is {"type": "email"}, which emails the monitor owner. Any extra field, such as an address, is rejected.

{"type": "email"}

Data Returned

Returns the monitor detail as structuredContent. The same detail is returned by bigdata_edit_monitor and by bigdata_get_monitor with an id. The result also comes as markdown: an overview table, JSON blocks for the search queries and the structured output schema, then the extraction instructions if they are set.

Field	Description
`id`	Monitor ID. Use it with the other monitor tools
`name`	Display name
`intent`	The intent the monitor was created from
`status`	`active` or `inactive`. Always `inactive` after create
`search_queries`	Stored fast-mode search payloads
`schedule`	`frequency`, `cron`, and `timezone`. Unused fields are `null`
`structured_output`	Field map, or `null`
`extraction_instructions`	Extraction instructions, or `null`
`entities`	RavenPack entity IDs, or `null`
`delivery`	Delivery settings, or `null`
`created_at`	Creation time
`updated_at`	Last update time, or `null`

{
  "id": "b0253b18-0dd8-4a01-a9a4-f209bce91a26",
  "name": "Anthropic – Regulatory Watch",
  "intent": "Track regulatory scrutiny of Anthropic PBC: any government, regulator, legislator or competition authority action aimed at Anthropic. Capture new or escalating investigations, probes, inquiries and information requests (e.g. FTC, DOJ, SEC, CMA, European Commission, EU AI Office, national data protection authorities), antitrust or merger reviews of Anthropic deals and investments, enforcement actions, fines, consent orders, regulator-filed lawsuits, government designations or restrictions (e.g. procurement bans, supply-chain risk designations, export controls), congressional or parliamentary hearings and letters targeting Anthropic, and compliance obligations or findings under AI laws such as the EU AI Act. For each event extract: the regulator or government body, jurisdiction, type of action, status/stage, the issue at stake, potential penalties or consequences, and the date.",
  "status": "inactive",
  "search_queries": [
    {
      "search_mode": "fast",
      "query": {
        "text": null,
        "texts": [
          "A regulator has opened an investigation into Anthropic.",
          "The competition authority is reviewing Anthropic's investment or acquisition deal.",
          "Lawmakers sent a letter to Anthropic or called it to testify at a hearing.",
          "A government agency designated Anthropic a supply-chain risk or restricted its use in federal procurement.",
          "Anthropic was fined or found in breach of the EU AI Act or data protection law."
        ],
        "auto_enrich_filters": null,
        "filters": null,
        "external_search": null,
        "max_chunks": null,
        "ranking_params": {
          "content_diversification": null,
          "source_boost": null,
          "freshness_boost": null,
          "reranker": {
            "enabled": false,
            "threshold": null,
            "text": null
          }
        }
      },
      "response_detail": "standard"
    }
  ],
  "schedule": {
    "frequency": null,
    "cron": "0 8 * * 1-5",
    "timezone": "Europe/Madrid"
  },
  "structured_output": {
    "regulator": {
      "description": "The proper name of the government body, regulator, legislature, committee or competition authority taking the action.",
      "field_type": "entity_reference",
      "values": null
    },
    "target_company": {
      "description": "The proper name of the company the action is aimed at, which is Anthropic PBC for this monitor.",
      "field_type": "entity_reference",
      "values": null
    },
    "action_type": {
      "description": "The kind of regulatory action taken against Anthropic.",
      "field_type": "enum",
      "values": [
        "investigation",
        "merger_review",
        "enforcement",
        "designation_restriction",
        "legislative_scrutiny",
        "compliance_obligation"
      ]
    },
    "status": {
      "description": "How far the action has progressed: opened when it is announced, filed or begun, concluded when it ends with a decision, outcome or taking effect; withdrawn if dropped without an outcome.",
      "field_type": "ordered_enum",
      "values": [
        "opened",
        "concluded",
        "withdrawn"
      ]
    },
    "jurisdiction": {
      "description": "The country or bloc whose authority the action comes under, as written in the source.",
      "field_type": "text",
      "values": null
    },
    "issue_at_stake": {
      "description": "The issue or conduct the action concerns, as written in the source.",
      "field_type": "text",
      "values": null
    },
    "potential_penalties": {
      "description": "Any potential penalties or consequences stated for Anthropic, as written in the source.",
      "field_type": "text",
      "values": null
    },
    "announcement_date": {
      "description": "When the change was announced or happened",
      "field_type": "datetime",
      "values": null,
      "read_only": true
    },
    "effective_date": {
      "description": "When it takes or took effect",
      "field_type": "datetime",
      "values": null,
      "read_only": true
    },
    "summary": {
      "description": "A short summary of the event, written by the system",
      "field_type": "text",
      "values": null,
      "read_only": true
    }
  },
  "extraction_instructions": null,
  "entities": [
    "IWGWUR"
  ],
  "delivery": null,
  "created_at": "2026-11-10T09:55:19.572327Z",
  "updated_at": "2026-11-10T09:55:19.683645Z"
}

Practical Tips

Pin only what you know

Start with just an intent. Pin search_queries, entities, or schedule only when the user has been specific, so the rest can be drafted from the intent.

Preview before activating

A new monitor does not run on a schedule or send email. Run a flash preview and read the results before turning it on.

MCP Reference

Tools

Official MCP Integrations

Custom MCP Integrations

Plugins

Prompt examples

Evals

Overview

When to Use

How It Works

Parameters

Important Notes

Shared Shapes

Schedule

Search queries

Entities

Structured output

Delivery

Data Returned

Practical Tips

Pin only what you know

Preview before activating

​Overview

​When to Use

​How It Works

​Parameters

​Important Notes

​Shared Shapes

​Schedule

​Search queries

​Entities

​Structured output

​Delivery

​Data Returned

​Practical Tips

​Pin only what you know

​Preview before activating

Overview

When to Use

How It Works

Parameters

Important Notes

Shared Shapes

Schedule

Search queries

Entities

Structured output

Delivery

Data Returned

Practical Tips

Pin only what you know

Preview before activating