salary-research

Pull live salary data for a role and location via the Apify MCP server and return a structured compensation summary.

Preflight check

Before doing anything else, confirm Apify MCP tools are available in this session (e.g. Call Actor, Get Actor run, Get dataset items). They appear when the user adds the Apify Desktop connector in Claude Customize → Connectors, pastes their token in the connector UI (not in chat), and sets Enabled tools to:

call-actor,get-actor-run,get-dataset-items,cheapget/best-job-search

If those tools are not available:

"Salary benchmarking needs the Apify Desktop connector. In Claude Desktop: Customize → Connectors → Desktop → Apify → Configure — add your Apify token, set Enabled tools to call-actor,get-actor-run,get-dataset-items,cheapget/best-job-search, save, enable the connector, then open a new chat and try again. Or run /career-navigator:launch Step 4 for the full walkthrough."

Stop here — do not proceed without the MCP connection.

Workflow

1. Collect inputs

Extract the following from the conversation. Do not ask for anything already stated:

Input	Source	Required
keyword	Job title from conversation or ## Target Roles in profile.md	Required
location	Location from conversation or ## Location in profile.md	Required
remoteOnly	Mentioned by user ("remote only", "fully remote")	Optional
minSalary	Mentioned by user ("above $120k", "minimum $150k")	Optional

If keyword or location cannot be determined from context or profile, ask for them together in a single prompt:

"What role and location should I pull salary data for? (e.g., 'Senior Product Manager in Chicago')"

2. Call the Apify actor

Use the Apify MCP call-actor tool with the following parameters:

• Actor ID: cheapget/best-job-search
• Input:

{
  "keyword": "<role title>",
  "location": "<city, state or country>",
  "maxResults": 25,
  "remoteOnly": <true if specified, omit otherwise>,
  "minSalary": <integer if specified, omit otherwise>
}

3. Parse the response

From each result record, extract these fields:

Field	Source key
Title	title
Company	company_name
Location	location
Remote	is_remote
Salary min	salary_minimum
Salary max	salary_maximum
Currency	salary_currency
Platform	platform
URL	platform_url
Posted	posted_date

4. Deduplicate

Where the same title + company_name combination appears across multiple location variants, keep only the record with the highest salary_maximum. If salary_maximum is null on all variants, keep the record with the highest salary_minimum. If all salary fields are null across all variants, keep the first occurrence.

5. Filter

Remove records where both salary_minimum and salary_maximum are null — unless doing so would leave fewer than 5 results. If fewer than 5 salary-bearing records remain after filtering, include all deduplicated records (with and without salary) so the table is still useful.

6. Sort

Sort the remaining records by salary_maximum descending. Records with no salary_maximum sort to the bottom, ordered by salary_minimum descending.

7. Output

Salary listings table

## Salary Benchmarks — {keyword} · {location}
Data via Apify · {today's date} · {n} listings

| Title | Company | Location | Remote | Min | Max | Platform |
|---|---|---|---|---|---|---|
| [title](platform_url) | company_name | location | ✓ / — | $salary_minimum | $salary_maximum | platform |

• Format salary values as $XXXk (e.g., $145k) or $XXX,XXX for precision if the source provides it. Use the salary_currency field — if not USD, show the currency code (e.g., £85k).
• If salary_minimum or salary_maximum is null for a record, show —.
• If is_remote is true, show ✓; otherwise show —.
• Embed platform_url in the title as a hyperlink.
• Show posted_date if available as a relative label (e.g., "3d ago") — omit the column if no records have it.

Compensation range summary

Below the table, add a summary block calculated across only the records that have at least one salary field populated:

### Compensation Range
- **Low:** $XXXk  (lowest salary_minimum across results)
- **Median:** $XXXk  (median of salary_maximum values, or salary_minimum where max is null)
- **High:** $XXXk  (highest salary_maximum across results)
- **Currency:** USD (or note if mixed)
- **Based on:** {n} of {total} listings with salary data

If fewer than 3 salary-bearing records exist, add a caveat:

"⚠️ Only {n} listing(s) included salary data — this range may not be representative."

Contextual note

If the user asked about a specific offer or their own comp floor, add one sentence comparing it to the range:

"Your offer of $135k base falls at the 60th percentile of listed ranges for this role and market."

Or if comparing to their profile's compensation floor:

"Your comp floor of $120k is below the median listed range — you have room to negotiate."

Only include this if there is a specific figure to compare against. Do not fabricate a comparison.

8. Error handling

Actor returns no results:

"Apify returned no salary data for '{keyword}' in '{location}'. Try a broader title (e.g., 'Product Manager' instead of 'Senior AI Product Manager') or a larger metro area."

All results filtered out (no salary data in any record):

"Found {n} listings for '{keyword}' in '{location}', but none included salary information. This is common for roles where compensation is listed as 'competitive' or 'DOE'. Try adding a minSalary filter to surface listings that disclose pay."

MCP call fails or times out:

"The Apify request didn't complete. In Claude Desktop, open Customize → Connectors → Apify, confirm your token and Enabled tools (call-actor,get-actor-run,get-dataset-items,cheapget/best-job-search), save, enable the connector, and try again in a new chat. See /career-navigator:launch Step 4 if you need the full walkthrough."