Data & Tools
Connect your AI assistant to live automotive data via Model Context Protocol. Standalone product — sign up just for MCP, or bundle with API access.
What It Does
One open standard. Every brand. New and used. The same data your dealer dashboard runs on.
Claude Desktop, Cursor, GPT-based agents, or your own custom client — if it speaks Model Context Protocol, it speaks VinAssessment. One connection gives your AI assistant access to live vehicle inventory, VIN history, competitor pricing, and market radius data. No bespoke integration required.
MCP is an open standard backed by Anthropic. Your AI assistant calls VinAssessment tools the same way it calls any other MCP tool — through the same configuration, the same auth flow, the same JSON responses. No proprietary adapter. No integration lock-in. When MCP clients improve, your VinAssessment connection improves with them.
Every tool call hits live data. Search inventory and get current listings — not a cached sample from last week. Look up a VIN and get its actual market position — priced against competitors within a radius you specify, updated continuously from 127M+ vehicle records and 2M new listings every day.
Connection Guides
One endpoint, one auth token, three clients we test against on every release.
macOS, Windows · stdio config
Step 1 — Edit your config
Open Settings → Developer → Edit Config and paste:
{
"mcpServers": {
"vinassessment": {
"command": "npx",
"args": ["-y", "@vinassessment/mcp"],
"env": {
"VA_API_KEY": "va_live_•••"
}
}
}
}
IDE · settings.json
Step 1 — Add MCP server
Open Cursor Settings → MCP Servers → New Server and configure:
# Server name vinassessment # Type command # Command npx -y @vinassessment/mcp # Env: VA_API_KEY va_live_•••
SSE / HTTP · any language
Step 1 — Hit the endpoint
Point any MCP-compatible client (or your own) at the SSE endpoint:
# Endpoint https://mcp.vinassessment.com/v1/sse # Auth header Authorization: Bearer va_live_••• # Tools exposed search_inventory · lookup_vin radius_analyze · price_band
Example Prompts
Three exchanges your AI assistant can have today, the moment you connect.
What Makes It Different
Five reasons builders pick our MCP Server over the alternatives — including not having one at all.
Sign up for MCP without an API subscription. You pay for the protocol connection you need, not the full API tier you don't.
Model Context Protocol is Anthropic-backed and open. Your integration follows MCP conventions, not a proprietary adapter that disappears when a vendor pivots.
Your AI assistant calls named, schema-defined tools (search_inventory, lookup_vin, price_band). Structured inputs, structured outputs. No prompt engineering required to extract useful data.
No cached snapshots returned to your assistant. Every tool call hits current listings and real-time pricing. The data your AI reasons about is the same data a dealer uses this morning.
If you also need REST API calls (batch VIN lookups, webhooks, raw data queries), bundle MCP with API Access. Or run MCP standalone. Both paths are supported.
Use Cases
Different jobs, same MCP server. The protocol meets your AI workflow where it lives.
You spend your day asking questions about the market — which trims are moving, which competitors just repriced, where a specific VIN sits in its radius. Instead of switching between your dashboard and your AI assistant, you ask Claude directly. The MCP Server connects Claude to live VinAssessment data, so your questions get answers from the same source as the dashboard — without the context-switching.
Your group has 12 rooftops and a custom Retool dashboard that your pricing managers use every morning. You want to add a natural-language query layer — "show me every vehicle on our lot priced above the 75th percentile in its radius." Connect your agent to VinAssessment via MCP, define the tools your Retool app can call, and your team gets answers in plain English backed by live data. No new dashboard to train them on.
You're building an agent that helps consumers or dealers navigate the car market — asking about price fairness, regional availability, or historical price trends. MCP gives your agent structured access to real listing data, competitor counts, and price bands without you writing and maintaining a bespoke API client. Your agent describes what it needs in natural language; the MCP tools return structured data your agent can reason about, cite, and display.
Tool Catalog
Twelve domains — VIN intelligence, dealer profiles, competitor pricing, auctions, market trends, and more. Every tool returns structured JSON with live data, scoped to your licensed perimeter.
lookup_vin
Full 360° report — decode, listings, price band, rank, title, auctions
decode_vin
Make, model, trim, engine, drivetrain, safety features from a VIN
decode_vin_batch
Bulk decode for up to 50 VINs in a single call
vin_history_listings
Full active and delisted listing history for a VIN
vin_price_history
Price time-series for chart timelines
vin_listing_count_by_state
Distinct US states the VIN has been listed in (proxy for migration)
vin_dealer_history
Distinct dealers that have ever listed this VIN, with date range
search_inventory
Filter by make, model, year, trim, price, mileage, ZIP + radius, body, fuel
search_inventory_by_text
Free-text search ("white pickup under $35k near Dallas")
find_similar_vehicles
Similar vehicles by ML feature vectors or visual matching
find_visually_similar_listings
Image similarity via perceptual hash (dhash / Hamming)
detect_duplicate_listings
Same VIN listed across multiple platforms — merged
search_new_arrivals
Listings added in the last N hours in a region
search_price_drops
Listings whose price dropped within a period
search_long_dom_listings
Stale inventory sitting more than N days
search_certified_preowned
CPO-only filter
search_trade_in_candidates
Vehicles matching a dealer's trade-in profile
my_inventory_overview
Caller's own inventory at a glance — units, mix, avg price, avg DOM
my_inventory_list
Full paginated list of caller's listings
dealer_inventory
Inventory for any authorized dealer by name or ID
dealer_profile
Name, address, phone, website, hours, photos, rating
dealer_reviews
Paginated Google reviews
dealer_search
Find dealers by name, ZIP + radius, or specialty make
dealer_inventory_age_buckets
DOM distribution by 0–30 / 31–60 / 60–90 / 90+ day buckets
dealer_pricing_position
Percent above or below market median
dealer_make_mix
Top makes by volume with share of stock
dealer_scraping_freshness
Last scrape time, vehicles added, vehicles removed
find_competitors
Comparable competitor vehicles for a VIN within a radius
competitor_count
Fast count of active competitors in a radius
competitor_rank
Caller's rank by price ("#12 of 47, below median")
competitor_price_distribution
Buckets — low, Q1, median, Q3, high — with histogram data
competitor_listings_table
Full competitor table with dealer, price, mileage, DOM, distance
competitor_price_change_alerts
Competitors who changed price (drops and raises)
competitor_new_listings
Net-new competitor listings in the last 24h or 7d
competitor_sold_listings
Listings that disappeared (proxy for sold)
dealer_vs_competitor_summary
Head-to-head summary across model, price, DOM
find_underpriced_competitors
Competitors significantly below market median
find_overpriced_competitors
Competitors significantly above market median
competitor_trims_breakdown
Trim-level gap analysis vs. caller's lot
price_band
Low, median, high for (year, make, model, trim) within a radius
price_band_by_zip
Same band aggregated per ZIP
price_band_with_mileage
Band that accounts for mileage bucket
suggest_price
Recommended price — three scenarios: aggressive, market, premium
trade_in_value
Fair / Good / Excellent trade-in estimate
price_trend_weekly
Weekly avg, min, max time-series
price_trend_monthly
Monthly aggregation with year-over-year change
mileage_price_correlation
Scatter data for odometer vs. price charts
dom_distribution
Days-on-market distribution by model or region
avg_dom_by_make
Median and mean DOM per make and model
price_vs_segment
Comparison to segment average (mid-size sedan, full-size SUV)
regional_price_comparison
Same model across multiple regions side-by-side
platform_price_comparison
Autotrader vs Cars.com vs Facebook vs CarGurus
calculate_loan_payment
Monthly payment, total interest, amortization
auction_lookup_vin
Every auction lot for a given VIN — hammer price, damage class, photos, sale date
auction_search
Search lots by year, make, model, state, damage, auction house
auction_price_band
Hammer-price low / median / high by damage class
auction_repair_cost_estimate
Average repair cost per damage class
auction_recent_sales
Most recent N sold lots for a model
auction_unsold_lots
No-sale lots and no-sale rate
damage_impact_on_value
How much a damage class drops price vs. clean
wholesale_to_retail_spread
Auction hammer vs retail listing for the same vehicle
listing_freshness_score
When last refreshed, days since price change, staleness flag
vin_market_footprint
Distinct platforms and dealers that have listed this VIN over time
spec_mismatch_check
VIN decode vs listing-claimed make/model/year disagreement
listing_completeness_score
Photos, description length, filled fields → quality grade A / B / C
fetch_window_sticker
Monroney sticker URL or PDF — generates if missing
window_sticker_summary
Structured fields — MSRP, options, packages, factory options
build_data_options
Factory options and safety features (HUD, leather, packages, ABS, ESC, blind-spot, lane-assist)
zip_neighbors
All ZIP codes within N miles
zip_to_geo
ZIP → city, state, lat, lon
state_inventory_summary
Listings per state with top makes
city_dealer_density
Dealer count and active vehicle count by city
map_aggregation
State → city → dealer → vehicle aggregation by zoom level
find_dealers_in_radius
Every dealer within a radius
market_supply_trend
Active-listing count for a model — weekly or monthly
market_demand_signal
Disappearance velocity (proxy sold-rate)
dealer_inventory_growth
How a competitor's stock changed over time
price_volatility_index
Price standard deviation over a period
seasonal_patterns
Seasonal rhythms — convertibles in spring, 4WD in winter
new_vs_used_share
New vs used listing share by region
listing_velocity
Average days from list → delist
vehicle_images
All image URLs for a VIN or listing
best_listing_photo
Cover or main image
photo_count_quality_check
Photo count plus a low-quality flag (under 5 photos)
vehicle_full_report
One-call markdown report — decode, history, prices, auctions, title, sticker
dealer_full_report
Profile + inventory + pricing position + competitor benchmark + reviews
market_report_by_model
Supply, price band, top dealers, regional split, trends
regional_market_report
Top makes, average price, DOM, hot models per region
price_change_report
Top-N listings with price changes over a period
inventory_aging_report
Caller's listings >60/90 days with new-price suggestions
gap_analysis_report
Trims competitors stock but caller does not
opportunity_report
Auction VINs with profitable retail flip spread
dealer_competitive_report
Head-to-head across 10 metrics
export_report
Markdown or CSV snapshot delivered as an MCP resource
Plus 4 MCP resources (catalog, data dictionary, dealer profile, glossary) and 5 prebuilt prompts (daily competitor brief, weekly market pulse, VIN deep dive, pricing review, auction opportunities) your AI client picks up automatically.
The Numbers
Vehicle records available through every tool call
New listings indexed every day — all current, no stale cache
Tools across 12 domains your AI assistant can call live
FAQ
lookup_vin
or
search_inventory
), receive a structured response, and reason about real data. MCP is open — any client that implements the protocol can connect to any compliant server, including VinAssessment's. Spec:
modelcontextprotocol.io
VA_API_KEY
environment variable for stdio transports (Claude Desktop, Cursor) or the
Authorization
header for SSE/HTTP. Each tool call sends the key with the request — no OAuth flow, no per-session token exchange. Rate limits apply per key. If you bundle MCP with API Access, the same key works for both.
Ready when you are
Connect it to 127M+ vehicles, real-time competitor pricing, and live inventory — in one MCP configuration. Standalone access available: no API subscription required.