com.claribi/mcp-server

Official26 tools

clariBI

AI business intelligence: sign up, ingest data, run analyses, generate reports from any MCP client.

Business intelligence platform for data ingestion, analysis, and report generation.

Captured live from the server via tools/list.

check_integration_status

Poll an OAuth handoff initiated by request_oauth_integration_url. Returns the current status (pending, connected, failed, expired) and, when connected, the data_source_id you can pass to run_analysis.

Parameters (1)

handoff_idstringrequired
The id returned by request_oauth_integration_url.

check_pricing

List clariBI subscription tiers with prices, AI credits, data source limits, user limits, and headline features. No authentication required.

Parameters (1)

tierstring
Optional. Return only this tier (free, trial, lite, starter, professional, enterprise).

create_checkout_session

Create a Stripe Checkout URL the user can open in a browser to upgrade their clariBI subscription. Payment cannot happen inside the LLM; this tool returns a URL.

Parameters (2)

tierstringrequired
Target subscription tier.
billing_periodstring
Billing cadence for the checkout. Defaults to monthly.

create_forecast

Bind a forecast to a metric in your workspace. Stores the configuration, sets up the schedule, and resolves the source binding immediately so a bad source_id or metric_path errors out before any credits are spent. Run the forecast with run_forecast.

Parameters (19)

namestringrequired
descriptionstring
source_typestringrequired
Which clariBI artifact the metric lives on. Most callers should start with data_source for raw connector metrics or goal to forecast a tracked KPI.
source_idstringrequired
UUID of the report, dashboard, goal, or data source.
metric_pathstringrequired
Dot-path to the numeric column. For data_source: the metric key (e.g. "revenue"). For goal: "value" or "percentage". For report: the path inside the insights JSON. For dashboard: "widget_id:metric_name".
granularitystring
Bucket size for the time series. Hourly requires an hourly data source; the engine refuses to fabricate hourly buckets from daily data.
horizon_daysinteger
How many BUCKETS ahead to project. The per-granularity cap applies: hourly 168, daily 90, weekly 52, monthly 24, quarterly 8, annual 5. The field name is "_days" for backwards compatibility — it counts buckets of the chosen granularity.
include_correlationsboolean
include_anomaliesboolean
include_changepointsboolean
narration_enabledboolean
When true, every completed run gets an AI-generated narrative attached (summary, highlights, risks, recommendations, methodology). Costs 1-2 extra AI credits per run.
aggregationstring
How multiple raw points falling into the same bucket are combined. Auto picks sum for additive metrics, mean otherwise. Use Mean for rates (CTR, conversion, latency).
transformstring
Series transform applied before fitting. Log helps revenue/traffic series with growing variance. Auto detects when log is beneficial; none forces raw scale.
non_negativeboolean
When true, point + lower band clamped at zero. Omit to let the engine heuristic decide (via non_negative_auto=true).
non_negative_autoboolean
When true (default), the backend heuristic owns non_negative and re-evaluates per run based on the metric path. Set false alongside non_negative to lock the choice.
method_overridestring
Force a specific forecasting method instead of auto-selecting via walk-forward CV. Valid names: naive, seasonal_naive, moving_average, drift, linear_trend, ar_p, holt_winters, holt_winters_multiplicative, holt_winters_damped, theta, gradient_boost, ensemble_top3. Empty = auto.
schedule_frequencystring
How often the forecast re-runs. Manual schedules only run when called explicitly via run_forecast.
schedule_day_of_weekinteger
schedule_day_of_monthinteger

generate_report

Create a new generated report in your clariBI organization. Returns the report id you can poll via get_report, plus a download URL once status reaches "completed".

Parameters (3)

titlestringrequired
Display title for the generated report.
template_idstring
Optional registry ID of the marketplace template to render. If omitted, an empty report shell is created and you can attach a template later from the web app.
output_formatstring
Output format. Must match one of ``GeneratedReport.OUTPUT_FORMATS``.

get_analysis_status

Check the status of a previously-dispatched run_analysis job. Returns the analysis result if completed.

Parameters (1)

job_idstringrequired
job_id returned by a prior run_analysis call.

get_billing_status

Get the organization's billing status — tier, renewal date, and upgrade options.

No parameters.

get_dashboard

Fetch one dashboard by ID. Includes widget definitions, the most recent refresh data, and the web URL.

Parameters (1)

dashboard_idstringrequired
UUID of the dashboard to fetch.

get_data_source_schema

Fetch the column schema for a data source. Useful before asking run_analysis about specific columns. The schema is derived from the preprocessing metadata clariBI extracted when the source was last synced. Poll this after upload_data_source / ingest_url_data_source until the returned status flips to "active" — that means preprocessing has finished and run_analysis will see the data.

Parameters (1)

data_source_idstringrequired
UUID of the data source to fetch the schema for.

get_forecast

Fetch one forecast configuration by ID.

Parameters (1)

forecast_idstringrequired
UUID of the forecast to fetch.

get_forecast_run

Fetch one run of a forecast by ID, or pass run_id="latest" for the most recent run. Returns the full forecast envelope: target series + 30-day projection, correlated drivers (with lag and bootstrap stability), anomalies, structural changes, and credit accounting.

Parameters (2)

forecast_idstringrequired
run_idstringrequired
UUID of the run. Pass 'latest' to fetch the most recent run for the forecast.

get_forecast_trust

Returns per-past-run accuracy for a forecast. Each row compares a prior run's projection against the actuals that have materialized since. Use this to see whether your forecasts have been getting MORE or LESS accurate over time, or to spot when a backtest sMAPE was systematically optimistic vs. realised performance.

Parameters (1)

forecast_idstringrequired

get_report

Fetch one generated report by ID.

Parameters (1)

report_idstringrequired
UUID of the report to fetch.

get_usage

Get the organization's current AI credit usage, data source count, user count, and rate-limit headroom.

No parameters.

ingest_url_data_source

Create a new data source by fetching a public URL on the server side. Handles CSV, TSV, JSON, Excel, TXT, and PDF. Private networks (RFC 1918, loopback, cloud metadata) are blocked. Returns the data_source_id once preprocessing has started. Use this for files larger than the 25 MB inline upload cap.

Parameters (5)

namestringrequired
Display name for the new data source.
urlstringrequired
Public http(s) URL of the file to ingest. The server fetches it once at call time; the URL is not re-fetched on subsequent analyses. Private networks (RFC 1918, loopback, link-local, cloud metadata) are blocked.
formatstring
Optional format hint. If omitted, the server infers it from the Content-Type header and URL extension.
descriptionstring
Optional human note stored on the data source.
wait_secondsinteger
Seconds to block waiting for preprocessing before returning. 0 returns immediately.

list_dashboards

List dashboards in your clariBI organization. Returns id, name, status, last refresh, and a URL you can open in a browser.

Parameters (3)

limitinteger
Page size (1–100).
offsetinteger
Row offset for pagination.
searchstring
Optional substring match against dashboard name.

list_data_sources

List the data sources connected to your clariBI organization. Returns id, name, source_type, status, last sync time, and the number of rows (when known).

Parameters (3)

limitinteger
Page size (1-100).
offsetinteger
Row offset for pagination.
source_typestring
Filter by source_type (csv, postgresql, google_ads, meta_ads, jira, mcp, …).

list_forecasts

List metric forecasts in your clariBI organization. Each row covers one metric: its source binding, horizon, schedule, and the latest run status. Use get_forecast_run to fetch the full forecast result.

Parameters (3)

limitinteger
Page size (1-100).
offsetinteger
Row offset for pagination.
is_activeboolean
Filter by active forecasts. Omit to include both active and paused.

list_reports

List generated reports in your clariBI organization. Returns id, title, status, output format, and download URL.

Parameters (3)

limitinteger
Page size (1-100).
offsetinteger
Row offset for pagination.
statusstring
Filter by status (pending, generating, completed, failed, cancelled).

regenerate_forecast_narrative

Re-run ONLY the AI narration step against an existing completed forecast run. Costs 1-2 AI credits (no engine work). Returns the new narration; the old one is overwritten in the run record. Refuses when the run is not yet complete or when narration_enabled=false on the forecast.

Parameters (2)

forecast_idstringrequired
run_idstringrequired
UUID of a COMPLETED run. Use get_forecast_run with run_id='latest' first if you want the most recent.

register_account

Begin clariBI account signup. Validates the email + organization name, emails a 6-digit verification code, and returns a pending_id. Call verify_email(pending_id, code) within 10 minutes to finish signup and receive an API key.

Parameters (5)

emailstringrequired
Work email for the new account.
organization_namestringrequired
Display name for the organization workspace.
first_namestring
User's first name (optional).
last_namestring
User's last name (optional).
accept_termsbooleanrequired
Must be true. By passing true the user agrees to https://claribi.com/terms and https://claribi.com/privacy.

request_oauth_integration_url

Initiate an OAuth handoff to a vendor integration (Google Ads, GA4, Search Console, Sheets, Drive, BigQuery, Meta Ads, Jira, Confluence). Returns an authorization URL the user opens in a browser. After the user clicks Allow, the connection is created and you can poll check_integration_status(handoff_id) to find out when the data is ready.

Parameters (2)

providerstringrequired
OAuth provider to authorize. Currently supports the native-OAuth catalog: Google (Ads, Analytics 4, Search Console, Sheets, Drive, BigQuery), Meta Ads, and Atlassian Jira / Confluence.
integration_typestring
Per-provider sub-type. Google accepts google_ads, gsheets, ga4, gsc, gdrive, bigquery, gcs, gcp, basic. Meta accepts ads or basic. Jira and Confluence accept basic. Defaults to basic.

run_analysis

Run a natural-language analytics question against your connected data sources. Consumes AI credits. Returns either the completed analysis result inline OR a job_id you can poll with get_analysis_status. If list_data_sources returns an empty list, ingest data first with upload_data_source (inline base64), ingest_url_data_source (public URL), or request_oauth_integration_url (Google / Meta / Jira / Confluence).

Parameters (3)

questionstringrequired
Natural language question to analyze. E.g. "What was revenue last quarter by region?".
session_idstring
Optional existing conversation session UUID.
wait_secondsinteger
How long (seconds) to wait for the job to finish before returning a job_id for polling. 0 = always return immediately.

run_forecast

Run a forecast now. Reserves AI credits up front, dispatches the backtest + projection + correlation pipeline, and returns either the completed result inline (wait_seconds > 0 and the run finishes in time) or a run_id you can poll with get_forecast_run.

Parameters (2)

forecast_idstringrequired
wait_secondsinteger
Seconds to wait for the run to finish before returning a poll handle. 0 = return immediately with the run_id.

upload_data_source

Create a new data source from an inline base64-encoded file (CSV, TSV, JSON, Excel, TXT, PDF). The file goes through the same validation and preprocessing as a web upload. Returns the data_source_id you can pass to run_analysis as soon as preprocessing completes (poll get_data_source_schema for readiness or pass wait_seconds to block here).

Parameters (5)

namestringrequired
Display name for the new data source. Shown in the web app and in subsequent list_data_sources calls.
formatstringrequired
File format. Drives MIME detection and the preprocessing route inside clariBI. csv covers comma-separated; tsv is tab-separated; xlsx is modern Excel; json must be a top-level array of objects or a single object.
data_base64stringrequired
Base64-encoded file contents. Maximum 25 MB encoded (~18 MB raw). For larger payloads, host the file at a public URL and use ingest_url_data_source.
descriptionstring
Optional human-readable description. Surfaces in the web app and in get_data_source_schema.
wait_secondsinteger
How long (seconds) to wait for preprocessing to finish before returning. 0 returns immediately with status="preprocessing"; the caller polls get_data_source_schema or list_data_sources.

verify_email

Complete clariBI signup by submitting the verification code plus a password. Returns an OAuth access_token for immediate use AND a long-lived MCP API key for persistent configuration. The new organization lands on the Trial tier (50 AI credits, 14 days).

Parameters (3)

pending_idstringrequired
The pending_id returned by register_account.
codestringrequired
6-digit verification code from the email.
passwordstringrequired
A password for the new account. Must be at least 8 characters and pass Django's standard validators.

Install

claude_desktop_config.json

{
  "mcpServers": {
    "mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://claribi.com/mcp/v1/"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.