Skip to content
RateSonar logo

RateSonar MCP server

Which MCP servers are RateSonar's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-31

← Back to RateSonar

Whose server it is
  • Registry namecom.ratesonar/ratesverified in the MCP registry

The registry name sits under ratesonar.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-31
  • Stateanswerschecked 2026-10-31
  • Tools11
  • Server calls itselfratesonar
  • Server version0.1.0
  • Protocol version2025-06-18

On 2026-10-31 the server answered initialize and tools/list without a sign-in. First answer in 66 ms.

Should an agent connect itNot proven either way yet
  • Whether its tools do the product's job has not been judged yet.
  • Checked on 3 of the 7 days it takes to judge reliability.

Once connected

  • Parameters are untyped or undescribed in 3 of 11 tools: check arguments before calling.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 3, the first on 2026-10-31
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 3 so far
  • Answeredmeasuring3 of the 7 daily checks it takes to print a share
  • Median first answer106 ms
  • Space in an agent's context≈ 3,213 tokensthe definitions of all its tools (11), which an agent carries on every turn
  • Tool definitionsA · 95 of 100
  • Typed results0 of 11 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 3 of 11

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers11 tools · read 2026-10-31
  • find_products

    Find current products by lender and words in the product name, e.g. institution "CBA" and query "standard variable". Returns institution_id and product_id for each match, with category, rate and observed_date. A product is the pair…

  • compare_rates

    Current comparison across institutions for a product category (home_loan | savings | term_deposit | credit_card | personal_loan | car_loan | business_loan): each product at its own most recent observation (observedDate tells you how…

  • rate_history

    Advertised rate history for one product (institutionId and productId). Returns each recorded rate, comparison rate and fees between two dates, with the date each was recorded. RateSonar's record starts 2026-06-12. Pass grid:true to also…

  • rate_details

    Exact CDR product-reference detail for one institution and product: every published lending or deposit rate entry and tier, with fees, features, eligibility, constraints, source quality flags, the date the version was recorded and the…

  • rate_changes

    Recent change-events, newest first: a rate moved, a product was added, or a product was withdrawn. Each event has the lender, the product's name and category, and field-level deltas (e.g. rate [6.09, 5.89], bps -20). Filter by…

  • list_series

    Every standard series slug the series tool accepts, with its label, kind (level or spread) and unit (% or bps). Call this before series when you do not know the slug.

  • series

    Standardized rate series: named LEVEL series (hl.var.oo.pi = headline variable home loan; sav.max; td.12m; biz.loan.var; cash = RBA cash rate…) and SPREAD series in bps (spread.hl.var-cash = lending pass-through; spread.hl.var-sav.max =…

  • bank_position

    Selected bank deposit and loan contracts against matching peer brands, own recorded history, RBA cash and dated funding disclosures. Pass an institutionId and product cohort. No claims about private ALM, intent or margins.

  • bank_analysis

    An evidence-checked AI reading of bank_position. Supports savings or term deposits against variable/fixed home loans, card purchase rates, personal, car or business loans. Returns cited claims, inference labels and dated records; partial…

  • yield_curve

    RBA curve nodes for an exact date or the latest publication: cash and bond yields as published, and the RBA's fitted zero, forward and discount curves (each point says whether it is observed or derived), with optional exact comparison…

  • curve_pipeline_status

    Validation, freshness, missing families, source hashes and last-good fallback status of the automated curve pipeline. An unhealthy response is evidence of a pipeline problem, not bank risk.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "rates": {
      "type": "http",
      "url": "https://api.ratesonar.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http ratesonar https://api.ratesonar.com/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://api.ratesonar.com/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.