MCP Tools

Overview

The MCP Server exposes five powerful tools that enable LLM agents to extract, analyze, and optimize SaaS pricing models. All tools accept pricing data either as a URL (for extraction) or as raw YAML content (for direct analysis).

All tool results are returned as JSON content blocks, eliminating the need for text parsing.

Common Parameters

These parameters are shared across multiple tools:

pricing_url

string

URL of the SaaS pricing page to extract and analyze. If provided, the MCP server will fetch and transform the pricing data via A-MINT.Example: https://buffer.com/pricing

pricing_yaml

string

Raw Pricing2Yaml YAML content to analyze directly, bypassing extraction. Useful when you already have structured pricing data.Note: Either pricing_url or pricing_yaml must be provided for all tools.

refresh

boolean

default:"false"

Force cache refresh. When true, bypasses cached pricing data and fetches fresh data from the source.

solver

string

default:"minizinc"

CSP solver to use for analysis. Options: minizinc, choco

minizinc: Open-source constraint modeling language
choco: Java constraint solver library

iPricing

Returns the canonical Pricing2Yaml (iPricing) document for a given pricing page or YAML content.

The tool name is iPricing (camelCase) in the MCP interface.

Parameters

pricing_url

string

URL of the SaaS pricing page to extract

pricing_yaml

string

Pre-existing Pricing2Yaml content to return

refresh

boolean

default:"false"

Bypass cache and fetch fresh pricing data

Response

request

object

Request parameters echoed back

url

string | null

The pricing URL provided (if any)

refresh

boolean

Cache refresh flag

pricing_yaml

string

Complete Pricing2Yaml document in YAML format

source

string

Source of the pricing data: amint (extracted) or upload (provided)

Example

await mcp_client.call_tool(
    "iPricing",
    pricing_url="https://github.com/pricing",
    refresh=False
)

Use Case

Use iPricing when you need to:

Extract structured pricing data from a URL
Retrieve the complete pricing model for manual inspection
Pass pricing data to other systems or tools

summary

Provides high-level statistics and metadata about a pricing model, including counts of plans, features, usage limits, and add-ons.

Parameters

pricing_url

string

URL of the SaaS pricing page

pricing_yaml

string

Pricing2Yaml content to summarize

refresh

boolean

default:"false"

Force fresh data extraction

Response

request

object

Request parameters

url

string | null

Pricing URL provided

refresh

boolean

Cache refresh flag

summary

object

Statistical summary of the pricing model

saasName

string

Name of the SaaS product

currency

string

Currency code (e.g., USD, EUR)

planCount

integer

Number of pricing plans

featureCount

integer

Total number of features

usageLimitCount

integer

Total number of usage limits

addOnCount

integer

Number of available add-ons

plans

array

List of plan names

features

array

List of feature names

usageLimits

array

List of usage limit names

Example

await mcp_client.call_tool(
    "summary",
    pricing_url="https://buffer.com/pricing"
)

Use Case

Use summary to:

Get a quick overview of a pricing model’s complexity
Validate that pricing data was extracted correctly
Understand the scope before running detailed analysis

subscriptions

Enumerates all valid subscription configurations within the pricing configuration space, optionally filtered by constraints.

Parameters

pricing_url

string

URL of the SaaS pricing page

pricing_yaml

string

Pricing2Yaml content to analyze

filters

object

Constraints to filter subscriptions. Structure:

{
  "features": [{"<feature_name>": <boolean_value>}],
  "usageLimits": [{"<limit_name>": <numeric_value>}],
  "addOns": ["<addon_name>", ...]
}

solver

string

default:"minizinc"

CSP solver: minizinc or choco

refresh

boolean

default:"false"

Force cache refresh

Response

request

object

Request parameters

result

object

Analysis results from the CSP solver

cardinality

integer

Total number of valid configurations found

subscriptions

array

List of valid subscription configurations

plan

string

Plan name

addOns

array

List of add-on names included

features

object

Feature values in this configuration

usageLimits

object

Usage limit values in this configuration

monthlyPrice

number

Total monthly cost

annualPrice

number

Total annual cost (if available)

Example

await mcp_client.call_tool(
    "subscriptions",
    pricing_url="https://buffer.com/pricing",
    solver="minizinc"
)

Use Case

Use subscriptions to:

Enumerate all possible pricing combinations
Find configurations matching specific requirements
Understand the configuration space size
Compare multiple valid options

optimal

Computes the optimal subscription (cheapest or most expensive) that satisfies given constraints.

Parameters

pricing_url

string

URL of the SaaS pricing page

pricing_yaml

string

Pricing2Yaml content to analyze

filters

object

Required constraints for the subscription. Same structure as subscriptions.

objective

string

default:"minimize"

Optimization objective: minimize (cheapest) or maximize (most expensive)

solver

string

default:"minizinc"

CSP solver: minizinc or choco

refresh

boolean

default:"false"

Force cache refresh

Response

request

object

Request parameters including filters and objective

result

object

Optimal subscription configuration

plan

string

Optimal plan name

addOns

array

Required add-ons (if any)

features

object

Feature values in optimal configuration

usageLimits

object

Usage limit values in optimal configuration

monthlyPrice

number

Monthly cost of optimal configuration

annualPrice

number

Annual cost (if available)

Example

await mcp_client.call_tool(
    "optimal",
    pricing_url="https://buffer.com/pricing",
    filters={
        "usageLimits": [{"channels": 10}],
        "features": [{"analytics": True}]
    },
    objective="minimize",
    solver="minizinc"
)

Use Case

Use optimal to:

Find the cheapest plan meeting user requirements
Identify the most feature-rich configuration within budget
Answer “what’s the best plan for X users with Y features?”
Provide pricing recommendations

validate

Validates the pricing configuration against the selected CSP solver to ensure mathematical consistency.

Parameters

pricing_url

string

URL of the SaaS pricing page

pricing_yaml

string

Pricing2Yaml content to validate

solver

string

default:"minizinc"

CSP solver: minizinc or choco

refresh

boolean

default:"false"

Force cache refresh

Response

request

object

Request parameters

result

object

Validation results

valid

boolean

Whether the pricing model is valid

errors

array

List of validation errors (if invalid)

message

string

Error description

location

string

Location in YAML where error occurred

warnings

array

Non-critical warnings about the model

Example

await mcp_client.call_tool(
    "validate",
    pricing_url="https://buffer.com/pricing",
    solver="minizinc"
)

Use Case

Use validate to:

Check pricing model consistency before deployment
Identify configuration errors in YAML
Ensure add-on dependencies are satisfied
Verify feature/limit relationships

Error Handling

All tools follow consistent error patterns:

Missing Input

{
  "error": {
    "code": -32602,
    "message": "Either pricing_url or pricing_yaml must be provided for <tool_name>."
  }
}

Invalid Solver

{
  "error": {
    "code": -32602,
    "message": "solver must be either 'minizinc' or 'choco'."
  }
}

Invalid Objective

{
  "error": {
    "code": -32602,
    "message": "objective must be 'minimize' or 'maximize'."
  }
}

Service Failure

{
  "error": {
    "code": -32603,
    "message": "Analysis job failed: <specific error>"
  }
}

Implementation Reference

The tool implementations are located in:

Tool definitions: /mcp_server/src/pricing_mcp/mcp_server.py:39-223
Workflow orchestration: /mcp_server/src/pricing_mcp/workflows/pricing.py
A-MINT client: /mcp_server/src/pricing_mcp/clients/amint.py
Analysis client: /mcp_server/src/pricing_mcp/clients/analysis.py

Harvey API

MCP Server

Analysis API

A-MINT API

CSP Service

Overview

Common Parameters

iPricing

Parameters

Response

Example

Use Case

summary

Parameters

Response

Example

Use Case

subscriptions

Parameters

Response

Example

Use Case

optimal

Parameters

Response

Example

Use Case

validate

Parameters

Response

Example

Use Case

Error Handling

Missing Input

Invalid Solver

Invalid Objective

Service Failure

Implementation Reference

Build docs developers (and LLMs) love

Harvey API

MCP Server

Analysis API

A-MINT API

CSP Service

​Overview

​Common Parameters

​iPricing

​Parameters

​Response

​Example

​Use Case

​summary

​Parameters

​Response

​Example

​Use Case

​subscriptions

​Parameters

​Response

​Example

​Use Case

​optimal

​Parameters

​Response

​Example

​Use Case

​validate

​Parameters

​Response

​Example

​Use Case

​Error Handling

​Missing Input

​Invalid Solver

​Invalid Objective

​Service Failure

​Implementation Reference

Build docs developers (and LLMs) love

Overview

Common Parameters

iPricing

Parameters

Response

Example

Use Case

summary

Parameters

Response

Example

Use Case

subscriptions

Parameters

Response

Example

Use Case

optimal

Parameters

Response

Example

Use Case

validate

Parameters

Response

Example

Use Case

Error Handling

Missing Input

Invalid Solver

Invalid Objective

Service Failure

Implementation Reference