Skip to main content
Glama
SynapCores MCP Server

SynapCores MCP Server

Official

Predict with a trained model

predict
Read-onlyIdempotent

Scores inline rows with a trained model, returning calibrated classification probabilities or regression values per row; use it for inference when all rows share the training columns.

Instructions

Score rows with a model trained by train_model and return {predictions, model_name, latency_ms, row_count}: calibrated probabilities in [0, 1] for classification (positive-class probability) or numbers for regression, one per input row. Use it for inference on rows you pass inline; use describe_model to check a model's metrics and features first. Every row must have the same columns, named like the training columns. Your tables are not modified.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "model_name": "<model_name>",
      "inputs": "<inputs>"
    },
    "name": "predict"
  }
}

Input Schema

NameRequiredDescriptionDefault
inputsYesRows to score, as objects mapping column names to values, e.g. [{"col1": 1, "col2": "a"}]; all rows need the same columns.
model_nameYesModel name, as registered by `train_model` or listed by `list_models`.
return_probaNoAccepted for compatibility; classification models always return calibrated probabilities.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changedv1.0.1
    • changedInput schema / properties / inputs / description
      Previous value: -"Array of feature dicts, e.g. [{col1: v1, col2: v2}, ...]. All rows must have the same columns."New value: +"Rows to score, as objects mapping column names to values, e.g. [{\"col1\": 1, \"col2\": \"a\"}]; all rows need the same columns."
    • changedInput schema / properties / model_name / description
      Previous value: -"Model name (the one passed to `train_model`)"New value: +"Model name, as registered by `train_model` or listed by `list_models`."
    • changedInput schema / properties / return_proba / description
      Previous value: -"Return probabilities (true, default) vs raw predictions (currently ignored - AIDB v1.6.5 always returns calibrated proba for classification)."New value: +"Accepted for compatibility; classification models always return calibrated probabilities."
  2. First observedv1.0.0

TDQS

A4.4/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnly/idempotent/non-destructive/closed-world, and the description adds real value beyond them: the exact return fields {predictions, model_name, latency_ms, row_count}, the semantics of the values (calibrated positive-class probabilities in [0,1] vs. regression numbers), and the rule that all rows must share columns named like the training columns. "Your tables are not modified" largely restates readOnlyHint, and error behavior for an unknown model name is not described.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the action and the return contract, then the usage routing, then the input constraint. Dense but zero filler; the closing note that tables are untouched is the only mildly redundant sentence.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description correctly carries the return-value burden by enumerating the four response fields and their semantics, and it covers the key input constraint. It stops short of describing failure modes (unknown model, column mismatch) or cost/latency expectations for large inputs.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the schema already defines all three parameters and the baseline is 3. The description adds meaning on top: it explains that prediction outputs are calibrated probabilities per row and clarifies the quirky `return_proba` flag (accepted for compatibility, ignored in practice), which the schema alone would not convey.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ("Score rows with a model trained by `train_model`") and immediately defines the return payload shape and meaning. It distinguishes itself from `describe_model` and `train_model` by scope, so an agent can pick it without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly scopes usage to "inference on rows you pass inline" and points to `describe_model` as a prerequisite check for metrics and features. It lacks an explicit when-not clause (e.g. large batches or table-resident data where `query`/`execute` would be preferable), so it falls short of full routing guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.