RateZip, an independent US consumer rates platform that compares savings, deposit, mortgage and lending rates, has launched a live rates service inside ChatGPT, giving users access to current US savings, CD, mortgage and HELOC rates.
The new RateZip Deposit/Mortgage Rates service uses the open Model Context Protocol (MCP) to bring live consumer rate data into ChatGPT. Each figure includes the source and observation time, allowing users to see which institution published the rate and when it was observed.
The service is the first US-focused MCP service dedicated to consumer deposit-rate comparison. The company said the service provides a source of truth for bank rates by returning the rate an institution currently publishes rather than relying on training data that may be months stale.
Users can ask ChatGPT questions such as which high-yield savings account currently offers the best rate, what current 30-year mortgage rates are, or whether a loan amount is conforming or jumbo. The service returns interactive rate cards that are sortable, timestamped and linked to the institution that published the figure.
Coverage spans high-yield savings accounts, certificates of deposit, mortgage rates and annual percentage rates, including conforming and jumbo loans, as well as home equity loans and HELOCs.
RateZip also allows users to enter a balance, which the service uses to resolve the relevant account tier and show the annual percentage yield a saver would actually earn rather than a headline rate that only applies above a certain threshold.
According to RateZip’s rate index, as of 26 August 2026, a handful of banks were still offering around 4.00% APY, compared with the Federal Deposit Insurance Corporation’s national average of 0.38%. RateZip said the difference is roughly $905 a year on a $25,000 balance, excluding taxes, balance tiers and future rate changes.
RateZip founder Paul Knag said, “Consumers are starting to ask AI assistants what to do with their money, and they deserve a factual answer with a timestamp and a source, not a guess from stale training data. Bringing real, checkable rates to ChatGPT puts an honest answer in front of the largest AI audience in the world.”
The service is built on the open MCP and can be called by any MCP-compatible agent through RateZip’s endpoint. It is also available as a custom connector in Perplexity, Grok and the Gemini CLI.
The development provides an example of how live, sourced financial data can be connected to AI assistants, a model that could be relevant to InsurTech as insurers and consumers explore AI-driven approaches to researching and comparing financial products.
RateZip said its data is factual information rather than financial advice, while the app carries no offers or advertising. Lender-specific quotes, approvals and account opening remain with the institutions, and rates and APYs can change without notice.
Copyright © 2026 InsurTech Analyst
