Docs / Pricing Advisor / MCP
Pricing Advisor MCP tools
Agents can inspect, explain, and apply Pricing Advisor recommendations via MCP. Human approval still fits in the loop.
Overview
The Pricing Advisor MCP tools let agents fetch pricing rules, list recommendations, explain why a change was suggested, and apply or reject it. This is useful for agent-native workflows that need to reason about pricing decisions.
When agents inspect
Agents use chargly.list_pricing_recommendations and chargly.get_pricing_recommendation to discover active recommendations. Use this when the agent needs to answer "What pricing changes are pending?" or "What does the current rule for image.generate look like?"
When they explain
chargly.explain_pricing_recommendation returns the reason, confidence, and estimated lift. Agents can surface this to users before any action — e.g. "Pricing Advisor suggests raising chat_message from 4 to 5 credits because margin is below target. Confidence: High. Estimated lift: +$182/month."
When they can apply or reject
chargly.apply_pricing_recommendation and chargly.reject_pricing_recommendation execute the decision. Apply creates a new immutable pricing rule version; reject preserves audit history. These tools require Growth plan and appropriate permissions.
Human approval in the loop
MCP does not bypass human control. Apply and reject are explicit actions — either you trigger them in the dashboard, or an agent does so in a workflow you authorize. There is no autonomous pricing. You can configure agent workflows to "explain only" and require manual apply in the dashboard, or allow the agent to apply when certain conditions are met (e.g. High confidence only).
MVP scope
No silent apply. Every apply/reject is auditable. Human approval still fits in the loop.
Tool reference
chargly.get_pricing_ruleFetch a pricing rule by feature or ID.chargly.list_pricing_recommendationsList recommendations for your project.chargly.get_pricing_recommendationGet a single recommendation by ID.chargly.explain_pricing_recommendationGet the reason and confidence for a recommendation.chargly.apply_pricing_recommendationApply a recommendation (creates new immutable version).chargly.reject_pricing_recommendationReject a recommendation (preserves audit history).See MCP Support for full tool documentation.