generated: '2026-07-21'
method: derived
status: candidate
source: https://docs.synthefy.com/
note: No official/hosted Synthefy MCP server was found (searched docs, npm @modelcontextprotocol, and
  MCP registries). This is a DERIVED candidate tool list mapped from Synthefy's two documented API operations,
  offered as a starting point for an agent integration — not a published server.
server:
  name: synthefy
  transport: http
tools:
- name: forecast_time_series
  description: Produce a multivariate time-series forecast from history and target dataframes using a
    Synthefy foundation model (e.g. Migas-1.0).
  source_operation: POST https://forecast.synthefy.com/v2/forecast
  inputs:
  - history_dfs
  - target_dfs
  - target_col
  - timestamp_col
  - metadata_cols
  - leak_cols
  - model
- name: nori_predict
  description: In-context tabular regression/classification with the Nori foundation model — predict targets
    for X_test given labeled X_train/y_train context.
  source_operation: POST https://inference.baseten.co/predict
  inputs:
  - model
  - task
  - X_train
  - y_train
  - X_test
deployment:
  mode: none
  verified: derived
  tools: 2
  checked: '2026-08-12'
  source: catalog MCP census