Tool reference

Estimated reading time: 10 minutes

Every tool, resource, and prompt the MCP server exposes today. The catalog is generated from running code at /api/mcp-server/tool-catalog/, so if you build clients against the live endpoint you always see the current set.

Public tools

No authentication required. Rate-limited to 5 requests per hour per IP in production.

register_accountpublic

Begin account signup. Validates the email and organization name, emails a 6-digit verification code, and returns a pending_id.

Arguments

Name	Type	Required	Description
`email`	string	Yes	Work email for the new account.
`organization_name`	string	Yes	Display name for the workspace.
`accept_terms`	boolean	Yes	Must be true. Accepts the terms of service and privacy policy.
`first_name`	string	No	Optional.
`last_name`	string	No	Optional.

verify_emailpublic

Finish signup by submitting the 6-digit verification code and choosing a password. Returns an access token and a long-lived MCP API key.

Arguments

Name	Type	Required	Description
`pending_id`	uuid	Yes	From `register_account`.
`code`	string	Yes	6-digit verification code from the email.
`password`	string	Yes	At least 8 characters.

check_pricingpublic

List clariBI subscription tiers with prices, AI credits, data source limits, user limits, and headline features.

Arguments

Name	Type	Required	Description
`tier`	string	No	Filter to one tier (trial, lite, starter, professional, enterprise).

Read tools

list_dashboardsdashboards:read

List dashboards in the organization. Returns id, name, status, last refresh, and a web URL.

Arguments

Name	Type	Required	Description
`limit`	integer	No	1 to 100, defaults to 20.
`offset`	integer	No	Defaults to 0.
`search`	string	No	Optional substring match against dashboard name.

get_dashboarddashboards:read

Fetch one dashboard by ID, including widget definitions and latest refresh data.

list_reportsreports:read

List generated reports with id, name, status, output format, and download URL.

get_reportreports:read

Fetch one generated report by ID.

list_data_sourcesdata-sources:read

List connected data sources with id, name, source_type, status, last sync time, and row count.

get_data_source_schemadata-sources:read

Fetch the column schema for a data source. Use this before asking run_analysis about specific columns.

get_usageusage:read

Current AI credit usage, data source count, user count, and rate-limit headroom.

get_billing_statusbilling:read

Subscription tier, renewal date, and upgrade URL.

Write tools

run_analysisanalysis:runtier: ai_analysis

Run a natural-language analytics question against connected data sources. Consumes AI credits. Returns the result inline if the analysis finishes within wait_seconds, or a job_id to poll.

Arguments

Name	Type	Required	Description
`question`	string	Yes	3 to 1000 characters.
`session_id`	uuid	No	Existing conversation session.
`wait_seconds`	integer	No	0 to 60, defaults to 30.

get_analysis_statusanalysis:run

Poll the status of a previously dispatched run_analysis job.

generate_reportreports:writetier: report_generation

Queue a new generated report. Returns a report_id and a polling URL.

Arguments

Name	Type	Required	Description
`name`	string	Yes	Report name, 2 to 200 characters.
`template_id`	string	No	Optional report template ID.
`output_format`	enum	No	pdf, docx, or html. Defaults to pdf.

create_checkout_sessionbilling:write

Create a Stripe Checkout URL the user opens in a browser to upgrade the subscription.

Arguments

Name	Type	Required	Description
`tier`	enum	Yes	starter, professional, or enterprise.
`billing_period`	enum	No	monthly or yearly. Defaults to monthly.

Resources

MCP resources are addressable by URI and read via resources/read. clariBI exposes three URI schemes:

Scheme	Pattern	Scopes
`dashboard`	`dashboard://<id>`	`dashboards:read`
`report`	`report://<id>`	`reports:read`
`data-source`	`data-source://<id>/schema`	`data-sources:read`

Prompts

Name	Description	Arguments
`analyze_question`	Build a well-formed analytics prompt around a user question, with the connected data sources as context.	`question` (string)
`weekly_review`	Kick off a weekly review of connected data sources, with anomaly flags and follow-up question suggestions.	(none)

Example: list dashboards

POST /mcp/v1/
Content-Type: application/json
Authorization: Bearer claribi_mcp_...

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "list_dashboards",
    "arguments": {"limit": 5}
  }
}

Response:

{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [{"type": "text", "text": "3 dashboard(s). Showing 3 starting at offset 0."}],
    "isError": false,
    "structuredContent": {
      "total": 3,
      "items": [
        {"id": "...", "name": "Sales Pipeline", "status": "active", "web_url": "https://claribi.com/app/dashboards/..."},
        {"id": "...", "name": "Marketing Funnel", "status": "active", "web_url": "https://claribi.com/app/dashboards/..."}
      ]
    }
  }
}

Build clients against the live catalog

The shipped Python and TypeScript SDKs generate their typed method surface from /api/mcp-server/tool-catalog/ at build time. New tools added on the server appear after re-running codegen.