Integration

CarsXE integration

CarsXE is a B2B vehicle data API: decode US and international VINs into specifications, decode license plates in 60+ countries, read a VIN or a plate number from a photo, look up market values, vehicle history (title brands, junk, salvage and insurance records), lien and theft records, open safety recalls (one VIN, a year make model, a manufacturer campaign code, or batches of up to 10,000 VINs), vehicle images, year make model data and dropdown options, OBD-II trouble codes, and a lender collateral score. Every successful lookup uses credits from the account's plan (1 to 50 per call, failed and no-data lookups are free). The ownership lookups (owner name and address by VIN, person, address or ZIP) are not exposed because they return personal data, nor the XML response format or the wallet (x402) payment flow. Authentication is a CarsXE API key from the dashboard, sent as the key query parameter.

Connect CarsXE How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in CarsXE.Built from the 25 CarsXE tools below, on the connection you add once.

What you can do with CarsXE

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

License plate, 4 operations

decode_plateDecode plateGETAgents: ask firstDecodes a license plate into a VIN and vehicle data in 60+ countries (country is an ISO code, optional for the US). For the US with no state it searches every state and answers results, count and billed_credits.
decode_plate_v1Decode plate v1GETAgents: ask firstDeprecated by CarsXE and due to be removed: prefer decode_plate. Decodes a plate into a VIN and vehicle data; state takes a US state, an Australian state, or one of the codes BR, UK, IR, IT, ES, IND, MX, NZ. 5 credits.
decode_us_plateDecode us plateGETAgents: ask firstLooks up a US vehicle by plate and state: VIN, year, make, model, trim, engine, body style, transmission, drive and fuel type, color. 5 credits; decodeVIN=true adds the full specifications for one more Specifications charge.
get_plate_from_imageGet plate from imagePOSTAgents: ask firstReads the license plate number from a photo of a vehicle (given as an image URL) and returns each plate found with its region and bounding box. Uses 2 credits per successful lookup.

Recall batch, 4 operations

download_recalls_batch_csvDownload recalls batch csvGETAgents: ask firstReturns the results of a completed or partial recalls batch as CSV text, one row per VIN. Answers an error while the batch is still running.
get_recalls_batch_resultsGet recalls batch resultsGETAgents: ask firstReturns the per-VIN recall results of a completed or partial recalls batch: for each VIN hasRecalls, recallCount and the recalls. Answers an error while the batch is still running.
get_recalls_batch_statusGet recalls batch statusGETAgents: ask firstReturns the status of a recalls batch from submit_recalls_batch: uploading, processing, completed, partial or failed, with VIN counts and the safety-recall hit rate. Poll every 30 to 60 seconds.
submit_recalls_batchSubmit recalls batchPOSTAgents: ask firstStarts a billed bulk recall check of up to 10,000 VINs (2 credits per unique VIN, charged at submission). Returns data.batchId: poll get_recalls_batch_status, then read get_recalls_batch_results.

Vehicle, 3 operations

decode_international_vinDecode international vinGETAgents: ask firstDecodes a non-US (international) VIN into vehicle attributes such as make, model and year. Uses 5 credits per successful lookup; for a US VIN use decode_vin.
decode_vinDecode vinGETAgents: ask firstDecodes a 17-character VIN into vehicle specifications: attributes (year, make, model, trim, engine, sizes), colors, equipment and warranties. Uses 2 credits per successful lookup.
get_vehicle_by_ymmGet vehicle by ymmGETAgents: ask firstLooks up vehicle data by year, make and model (optionally trim), no VIN needed: the best matching vehicle and, with allTrimOptions=1, every trim offered. Uses 2 credits per successful lookup.

Collateral score batch, 2 operations

get_collateral_score_batchGet collateral score batchGETAgents: ask firstReturns the progress of a collateral score batch (status, processed, succeeded, failed) and, once status is completed, every result in submission order.
submit_collateral_score_batchSubmit collateral score batchPOSTAgents: ask firstStarts a billed collateral score batch of up to 500 vehicles (for existing customers on legacy plans). Returns batch_id: poll get_collateral_score_batch until status is completed. One batch per account at a time.

Market value, 2 operations

get_market_valueGet market valueGETAgents: ask firstReturns detailed market values for a VIN: wholesale, retail and trade-in values per condition tier, auction values, mileage bands and option adds and deducts. Uses 25 credits per successful lookup.
get_market_value_v1Get market value v1GETAgents: ask firstLegacy market value lookup (v1) by VIN: retail, trade-in (clean, average, rough), loan value, MSRP, a trade-in history and auction values. Uses 25 credits; get_market_value (v2) adds condition tiers and mileage bands.

Recall, 2 operations

get_vehicle_recallsGet vehicle recallsGETAgents: ask firstReturns the open safety recalls for one vehicle by VIN (NHTSA data: campaign numbers, components, remedies). Uses 5 credits per successful lookup, also when no recall is found.
list_recalls_by_modelList recalls by modelGETAgents: ask firstLists the open safety recalls for a year, make and model, no VIN needed: recall_count, has_recalls and the recalls. Uses 1 credit per successful lookup, also when none is found.

Collateral score, 1 operation

get_collateral_scoreGet collateral scorePOSTAgents: ask firstScores a vehicle as loan collateral: mileage and region adjusted values, depreciation forecast at 12 to 48 months, title and odometer flags, open stop-sale recalls and a risk tier (A to D) with a suggested max LTV.

Lien theft record, 1 operation

get_lien_theft_recordsGet lien theft recordsGETAgents: ask firstChecks a VIN for active lien and theft records in the US and returns the vehicle (year, make, model, type) with the events found (event, location, details). Uses 5 credits per successful lookup.

Obd code, 1 operation

decode_obd_codeDecode obd codeGETAgents: ask firstDecodes an OBD-II diagnostic trouble code (for example P0115) into its diagnosis text. Uses 1 credit per successful lookup.

Recall campaign, 1 operation

get_recall_campaignGet recall campaignGETAgents: ask firstLooks up a manufacturer recall campaign code (for example Volkswagen 37M2) and returns the campaign: class, description, NHTSA numbers, affected models and years. An unknown code answers an error and uses no credits.

Vehicle history, 1 operation

get_vehicle_historyGet vehicle historyGETAgents: ask firstReturns a vehicle history report for a VIN: current title, title brands, junk and salvage records, insurance records and history events. Uses 25 credits; lienTheft=true adds a lien and theft check.

Vehicle image, 1 operation

search_vehicle_imagesSearch vehicle imagesGETAgents: ask firstFinds images of a vehicle by make and model (optionally year, trim, color, angle, interior or exterior, size, usage rights). Returns image links with sizes and the page each was found on. 2 credits.

Vehicle option, 1 operation

list_ymm_optionsList ymm optionsGETAgents: ask firstReturns one layer of year, make, model, trim or variant values for dropdown menus (years, makes, models, trims or variants), inferred from the filters or set by dimension. 1 credit, or 1 per model for bulk variants.

Vin, 1 operation

get_vin_from_imageGet vin from imagePOSTAgents: ask firstReads a VIN from a photo (an image URL or a base64 image string) and returns the VIN with its confidence, bounding box and other candidates. Uses 1 credit per successful lookup.

Use CarsXE from Claude Code, Cursor or Codex

Connect CarsXE in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its CarsXE operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect CarsXE

Create your workspace
Sign up free. Nothing to install.
Connect CarsXE once
Add your CarsXE connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.