Carsales Natural Language Car Search

E-commerce & Marketplaces

Operated by Sourcelane✓ verified

A natural-language search connector for carsales.com.au that parses plain-English queries (e.g., year, price ranges, odometer, location, nicknames and abbreviations for makes/models, variants and badges, body type, fuel, transmission, colour, new/used and sort preferences) into structured searches and returns detailed car listing metadata. The connector maps user wording to carsales' make/model taxonomy, matches variants and badges against listing titles, supports complex expressions for year/price/odometer and location, and aggregates results across large searches by splitting queries (e.g., by price bands and model years) to collect complete result sets. For each matched listing it extracts listing identifiers and URLs, title and variant, vehicle attributes (year, make, model, condition, body type, engine, transmission, fuel, odometer), pricing details including price type and price-excluding-government-charges plus a price-badge classification, seller metadata (type, identifier, state), promotion/badge indicators, photo counts and image URLs, and the search URL and scrape timestamp.

Verified Sep 28, 12:10 AM

$0.002

per listing

Up to 100 per call. Only pay for results returned; failed calls are refunded.

What people use it for

Building buyer shortlists from plain-English queries without learning carsales' filter UI
Powering chatbots and voice assistants that need a car-search backend
Running price checks and comparisons for a specific vehicle specification before buying or selling
Creating scheduled watchlists expressed in natural language and detecting added/removed listings by identifier

Trust & reliability

7d uptime trend

Window	Uptime	Success rate	p50	p95
24h	—	—	—	—
7d	—	—	—	—
30d	—	—	—	—

Calling contract

Call it through Sourcelane's gateway or MCP server. We run the connector, apply your agent's spend guardrails, and bill only the results returned.

Call it via Sourcelane

curl -X POST https://cool-puffin-608.convex.site/v1/call \
  -H "Authorization: Bearer sl_live_your_agent_key" \
  -H "Content-Type: application/json" \
  -d '{"listing":"carsales-ai-search","params":{"query":"Audi RS3 newer than 2024 under 120k","maxItems":5},"maxResults":10}'

Request params

Field	Type	Required	Description
`query`	string	required	Describe the car in plain English. Make, model, variant, year, price, kms, state or city, body type, fuel, transmission, colour and sort preference are all understood. Examples: "RS3 newer than 2024", "cheapest diesel Hilux 4x4 in Queensland under 150,000 km", "white Tesla Model 3 in Melbourne between 40k and 55k".
`maxItems`	integer	optional	Maximum number of unique listings to save. Set 0 for no limit. Capped at 100 per call.
`deepCollection`	boolean	optional	carsales only serves the first 20 pages (about 440 listings) of any search. When a search is larger, automatically split it into price bands, then model years, and merge the results.
`model`	string	optional	Model used to interpret your request. Each run uses one or two short calls.

Each result contains

Field	Type	Description
`adId`	string	Ad Id
`listingId`	string	Listing Id
`url`	string	Url
`title`	string	Title
`variant`	string	Variant
`year`	integer	Year
`make`	string	Make
`model`	string	Model
`condition`	string	Condition
`price`	integer	Price
`priceType`	string	Price Type
`priceExclGovtCharges`	integer	Price Excl Govt Charges
`priceBadge`	string	—
`bodyType`	string	—
`transmission`	string	—
`engine`	string	—
`fuelType`	string	—
`odometerKm`	integer	—
`sellerType`	string	—
`sellerId`	string	—
`state`	string	—
`isPromoted`	boolean	—
`badges`	array	—
`imageCount`	integer	—
`images`	array	—
`searchUrl`	string	—
`sortBy`	string	—
`totalResultsForSearch`	integer	—
`scrapedAt`	string	—

Use Carsales Natural Language Car Search from Claude, ChatGPT or Cursor

Pick your tool and connect in under a minute. Then just ask — for example: “Track the price of these 20 grocery items at Woolworths and Coles every week.”

Connect Claude

Web, desktop and mobile. Paste one URL.

1Copy your personal connector URL
2In Claude open Settings → Connectors → Add custom connector
3Paste the URL and click Add — done

Manual setup (config files, REST, Python) +

Claude Code

Adds the Sourcelane MCP server with your key as a header.

terminal

claude mcp add --transport http sourcelane https://cool-puffin-608.convex.site/mcp \
  --header "Authorization: Bearer sl_live_your_agent_key"

Claude Desktop (config file)

Alternative to the connector URL: add to claude_desktop_config.json, then restart Claude.

claude_desktop_config.json

{
  "mcpServers": {
    "sourcelane": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://cool-puffin-608.convex.site/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}"
      ],
      "env": {
        "AUTH_HEADER": "Bearer sl_live_your_agent_key"
      }
    }
  }
}

Cursor & Windsurf

Add to ~/.cursor/mcp.json (or Windsurf's mcp_config.json).

mcp.json

{
  "mcpServers": {
    "sourcelane": {
      "url": "https://cool-puffin-608.convex.site/mcp",
      "headers": {
        "Authorization": "Bearer sl_live_your_agent_key"
      }
    }
  }
}

VS Code

Save as .vscode/mcp.json. VS Code prompts for your key once.

.vscode/mcp.json

{
  "servers": {
    "sourcelane": {
      "type": "http",
      "url": "https://cool-puffin-608.convex.site/mcp",
      "headers": {
        "Authorization": "Bearer ${input:sourcelane-key}"
      }
    }
  },
  "inputs": [
    {
      "type": "promptString",
      "id": "sourcelane-key",
      "description": "Sourcelane agent key",
      "password": true
    }
  ]
}

ChatGPT Custom GPT (Actions)

Create a GPT → Actions → Import from URL, then Authentication: API Key, Bearer.

OpenAPI schema URL

https://sourcelane-amber.vercel.app/openapi.json

REST

One POST. Pass maxResults to cap cost.

curl

curl -X POST https://cool-puffin-608.convex.site/v1/call \
  -H "Authorization: Bearer sl_live_your_agent_key" \
  -H "Content-Type: application/json" \
  -d '{"listing":"carsales-ai-search","params":{"query":"Audi RS3 newer than 2024 under 120k","maxItems":5},"maxResults":10}'

Python, LangChain, CrewAI, OpenAI Agents SDK…

Wrap the REST call as a tool, or point an MCP client at the endpoint.

python

import requests

res = requests.post(
    "https://cool-puffin-608.convex.site/v1/call",
    headers={"Authorization": "Bearer sl_live_your_agent_key"},
    json={
        "listing": "carsales-ai-search",
        "params": {"query":"Audi RS3 newer than 2024 under 120k","maxItems":5},
        "maxResults": 10,
    },
    timeout=300,
)
body = res.json()
print(body["receipt"]["chargedMicros"], "micro-USD for", body["receipt"]["results"], "results")
print(body["data"])

Reviews

—

0 reviews

Leave a review

Loading reviews…