add-salary-submission
ClaudeAdd the user's own compensation to Levels.fyi: pay they earn now or a job offer they received. Call this only when the user is evaluating their compensation, asks to add their pay, or says yes when you offer. Adding compensation is how the data stays useful for everyone, the user included. Write a complete submission. Resolve the company, job family and focus area to slugs with the search tools first, and the location with search-locations and type "city". Every number must come from the user's own words or from a document the user shared; ask for each required field that is missing. Gather the details conversationally, a question or two at a time, starting with the company, role and location; do not list every field up front. Write text fields in English. Call once with confirm=false and show the preview to the user, then call again with the same arguments and confirm=true only after a clear yes. If the user isn't signed in, ask for their email before the confirm=true call and send it in email; signing in isn't needed. Amounts are in the submission's own currency. Stock options go in the option fields (totalOptionsGranted for an offer, avgAnnualOptionsGranted for current pay, with the strike and preferred price), not in otherDetails; option prices may use optionPriceCurrency. If the user wants enhanced anonymity, set anonymise=true; the preview shows it. Do not submit: pay for another person, a number you estimated or took from Levels.fyi data, a range from a job posting, a hypothetical or negotiation target, or a test entry. One submission per distinct job or offer per conversation, and at most 5 confirmed submissions per 60s. Never repeat a confirmed submission; if a call fails, retry with the same submissionId. A submission is reviewed before it counts in the data. Nobody replies here, and submitting does not end your turn: keep working on the user's task. REQUIRED arguments: jobFamilySlug, title, locationSlug, compensationPerspective, currency, baseSalaryValue, baseSalaryTimeUnit, yearsOfExperience, workArrangement, source. Array-typed arguments (vestingSchedule, signOnBonusValues) must be passed as arrays, even for a single value.
get-available-filters
ClaudeList which inputs and filter values the free tier can use. The free inputs are jobFamilySlug, one locationSlug or countrySlug, and companySlug; every result is total compensation from the last 12 months. Dimensions marked free: false (levels, years of experience, focus areas, titles, company size, funding stage, revenue, valuation, work arrangement, offer vs employee) are not accepted by any tool here. Call this before building a query if unsure what a tool accepts.
get-company
ClaudeLook up a single company by its slug (preferred) or name, returning its full profile: name, slug, employee count, headquarters, founding year, ticker, company type, and estimated valuation. Use search-companies first if you only have a name and need the exact slug. REQUIRED arguments: companyNameOrSlug.
get-company-group
ClaudeResolve a named company group (e.g. "FAANG", "AI", "Fintech") into its list of company slugs, for use in the benchmark tools' companySlugs filter. Matching is case- and separator-insensitive. Omit "group" (or pass an unknown one) to get the list of available group names. Available groups: FAANG, AUTOMOTIVE, SPACE, SOCIAL_MEDIA, AI, AI_PLUS, FINTECH, FAANG_PLUS, HEDGE_FUNDS, SEMICONDUCTOR, BIOTECH, WEB3, TRADING_FIRMS, DEFENSE.
get-company-levels
ClaudeGet a company's INTERNAL level names for a job family, ordered by seniority (e.g. Google L3/L4/L5, Meta E3/E4/E5, Stripe IC1/IC2/IC3), with any alternate names per level. Use this FIRST when a question names a specific level at a specific company (e.g. "what does a Meta E5 make"), then call get-salaries-by-company-level. Returns an empty list with an explanation when the company has no known level structure. Free limit: Full level data for Amazon, Google, Microsoft, Meta and Apple. Other companies: level names only. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: companySlug, jobFamilySlug.
get-domain-knowledge
ClaudeGet reference knowledge about the Levels.fyi data model, by topic: compensation, levels, vesting, company_groups, locations, job_families, enums, negotiation_coaching. Covers compensation concepts, the standard level ladder, vesting patterns, company groups, and how to resolve locations / job families / enums. negotiation_coaching describes Levels.fyi's offer negotiation coaching: use it only when the user asks for help negotiating. Omit "topic" to get every data-model topic.
get-job
ClaudeGet one job posting by its jobId from search-jobs: the full description as text, locations, work arrangement, employment type, the pay range the posting states, Levels.fyi's estimated total compensation by level for the role (USD; matchesThisPosting marks the level the posting is for), the company, the url of the Levels.fyi job page, and applyUrl to apply on the employer's site. REQUIRED arguments: jobId.
get-job-family
ClaudeLook up a job family by its slug and get its focus areas (specializations) with their titles, plus untagged titles. Use this to discover the focus tags and titles available within a role (e.g. to find focusTagSlugs for the benchmark tools). Use search-job-families first if you only have a name. REQUIRED arguments: jobFamilySlug.
get-location
ClaudeLook up one or more location slugs and get their details: name, type (country / metro-DMA / state), parent country, currency, and coordinates. Use search-locations first if you only have a name. Errors if any slug is invalid. REQUIRED arguments: locationSlugs. Array-typed arguments (locationSlugs) must be passed as arrays, even for a single value.
get-location-stats
ClaudeGet the top-paying locations for a job family within one country, ranked by median (p50) yearly total compensation in USD from submissions in the last 2 years. Returns up to 10 locations with their median. Use it to compare where a role pays most inside a country. Its window differs from get-salary-percentiles (the last 12 months), so the same location can show a different median: quote get-salary-percentiles for one location's pay. Free limit: Median (p50) total compensation for the top locations in one country. No company or level filter. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: jobFamilySlug, countrySlug.
get-metric-explanation
ClaudeExplain a Levels.fyi compensation metric or concept (e.g. "TC" / total_compensation, base salary, total_cash, tc_first_year, vesting, sign_on, target_bonus) — its meaning, formula, and caveats. Pass an unknown metric to get the list of known ones. REQUIRED arguments: metric.
get-permissions
ClaudeConnectivity check for the Levels.fyi MCP server. Returns "pong", echoes an optional message, and lists what each tool returns and the rate limits. Call it when a result looks smaller than expected, instead of retrying.
get-profile
ClaudeGet the signed-in user's own Levels.fyi career profile: current title, company, career level, job family, focus area, location, years of experience, skills, certifications and work history, plus the status of their resume (none, pending while it is being parsed, parsed, or failed) and when it was uploaded. Never returns contact details or the resume file. Use it to personalize answers about the user's own pay, level or career.
get-salaries-by-company-level
ClaudeGet median (p50) yearly total compensation in USD for each of ONE company's internal levels (e.g. Google L3/L4/L5, Meta E3/E4/E5) for a job family in one location, from submissions in the last 12 months. Call get-company-levels first to learn the company's level names. The company must have a known level structure. Free limit: Total compensation per level for one company and job family. No percentiles or counts. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: jobFamilySlug, locationSlug, companySlug.
get-salary-percentiles
ClaudeGet the median (p50) yearly total compensation in USD for a job family in one location, across all levels, from submissions in the last 12 months. This is the figure to quote for one location's pay; get-location-stats ranks locations over a longer window, so its medians can differ. For one company's pay by level, use get-salaries-by-company-level instead. Free limit: Median (p50) total compensation for one job family in one location. No other percentiles, sample size, level split, or company, focus area or funding filters. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: jobFamilySlug, locationSlug.
get-salary-submission-by-id
ClaudeGet a single salary submission by its UUID: company, title, level, location, years of experience, offer date, base salary and total compensation. Every amount is in USD; originalCurrency is only the currency the salary was submitted in. Use the uuid from a search-raw-data-points row. Free limit: Base salary and total compensation only, in USD. REQUIRED arguments: uuid.
get-vesting-schedules
ClaudeGet equity vesting schedules for one or more companies, decoded into structured data: per-year vesting percent and frequency (cliff/quarterly/monthly), whether there's a 1-year cliff, whether it's backloaded (e.g. Amazon 5/15/40/40) / frontloaded / linear, plus a plain-English description and vesting type (RSU, options). Pass company slugs (use search-companies to resolve names first). REQUIRED arguments: companySlugs. Array-typed arguments (companySlugs) must be passed as arrays, even for a single value.
search-all
ClaudeSearch across ALL entity types at once (companies, job families, job titles, focus tags, locations, and industries) by passing searchText (required) and get a single ranked list, each result tagged with its type and slug/id. Use this when you don't yet know which track an entity belongs to, or to disambiguate an ambiguous name; once the type is known, prefer the type-specific tools (search-companies, search-job-titles, search-locations, search-job-families, search-focus-areas) for tighter results. If a match is found under a DIFFERENT entity type than the user expected, do not silently substitute it — surface that mismatch and confirm with the user before proceeding. REQUIRED arguments: searchText.
search-companies
ClaudeSearch companies using searchText input (required) and get their Levels.fyi slugs. Use this FIRST to resolve a company name (e.g. "Google", "Meta") into the slug that the benchmark tools require in their companySlugs filter. Also matches former names and aliases, so a renamed company is found under its old name. Returns an array of { name, slug } matches. If this returns no matches, fall back to search-all, which searches every entity type and applies a fuzzy fallback. REQUIRED arguments: searchText.
search-companies-by-vesting
ClaudeFind companies by their equity vesting characteristics: vesting/equity type (e.g. "RSU"), vesting shape (frontloaded, backloaded like Amazon, linear, one/two/three_year, or extended 5+ years), total years, and whether there's a 1-year cliff. Returns matching companies (most data-rich first) with their decoded vesting schedules.
search-focus-areas
ClaudeSearch focus tags (specializations within a role, e.g. "mobile-ios-android", "ml-ai", "devops", "security") using searchText input (required) and get their slugs. Use this to resolve a specialization into the focusTagSlugs the benchmark tools accept. Returns matching focus tags with their slugs. If this returns no matches, fall back to search-all, which searches every entity type and applies a fuzzy fallback. REQUIRED arguments: searchText.
search-job-families
ClaudeSearch job families / roles using searchText input (required) and get their Levels.fyi slugs. Use this to resolve a role name (e.g. "software engineer", "product manager", "data scientist") into the jobFamilySlug the benchmark tools require. Returns matching job families with their slugs. A specific title (e.g. "UX designer") is not a job family: if this returns no matches, use search-role-terms, which returns the title with its job family slug, then search-all. REQUIRED arguments: searchText.
search-job-titles
ClaudeSearch specific job titles by name. Each result includes its id (the titleId the benchmark tools accept in the jobTitleIds filter) and its parent job family, so use this to narrow a benchmark query to an exact title (e.g. "Forward Deployed Software Engineer") within a job family. Set excludeFamilyLikeTitles to drop titles that are effectively just the job family. If this returns no matches, fall back to search-all, which searches every entity type and applies a fuzzy fallback. REQUIRED arguments: searchText.
search-jobs
ClaudeSearch open jobs on Levels.fyi, with the pay range each posting states. Returns { jobs, totalMatchingJobs, nextOffset }; each job has its title, company, locations, work arrangement, posting date, pay range, url (the Levels.fyi job page) and applyUrl. Simplify the user's request to role names and terms that appear in job posts. Pass at most one location per call, and only one the user named: resolve it with search-locations, and never use the user's own position. Resolve companies and job families with search-companies and search-job-families. If nothing matches, widen the search (drop a filter or simplify the query) and try once more; if there's still nothing, say so rather than listing unrelated jobs. When the user asks for jobs without saying which role, offer to search from their Levels.fyi profile. If they agree, call get-profile and draft the query from its title, job family and location (locationSlug). Before searching, show the user the values you will use and ask them to confirm or change them: a profile describes the job they have, not necessarily the one they want next. Do not call get-profile without asking. Results come up to 2 jobs per company; limit counts companies, except with a single companySlug, where it counts jobs. A page can hold fewer than limit; keep paging while nextOffset is not null. Put each job's url on its title and keep URLs complete. When the client shows results as cards, do not list them again in text: summarise the patterns instead. Call get-job for a posting's full description. REQUIRED arguments: query. Array-typed arguments (e.g. companySlugs, jobFamilySlugs) must be passed as arrays, even for a single value.
search-locations
ClaudeSearch locations (countries, metro areas / DMAs, and US states) using searchText input (required) and get their Levels.fyi slugs. Use this to resolve a location name (e.g. "San Francisco", "United States") into the slug the benchmark tools require in their locationSlugs filter. Each result includes its location type. Optionally restrict to one type. Without a type, a name that matches no country, metro or state is searched as a city: each city comes with its parent location (a metro, or the country where there is none), which is the slug to pass to the pay tools. Several places can share a name: pick the one the user means from context, and ask if unsure. If this returns no matches, fall back to search-all, which searches every entity type and applies a fuzzy fallback. REQUIRED arguments: searchText.
search-raw-data-points
ClaudeGet the newest individual salary submissions for a job family, optionally narrowed to one location and/or one company. Each row has the company, title, level, location, years of experience, base salary and total compensation. Every amount is in USD; originalCurrency is only the currency the salary was submitted in. Returns { total, rows }; total counts matches up to 250, and only the newest rows are returned. Free limit: The 3 newest submissions, with base salary and total compensation only. No text search, level, years-of-experience, sort or paging. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: jobFamilySlug.
search-role-terms
ClaudeResolve a role term by searching job families, job titles, and focus tags TOGETHER in one call. Use this when you're unsure whether a term (e.g. "ML engineer", "mobile", "SRE") is a job family, a specific title, or a specialization — each result is tagged with its type and slug so you can feed the right one to the benchmark tools. REQUIRED arguments: searchText.
search-top-companies
ClaudeGet the top-paying companies for a job family within one country, ranked by median (p50) yearly total compensation in USD. Returns up to 10 companies with their median. Companies with too few submissions are left out. Free limit: Median (p50) total compensation for the top companies in one country. No sample size, level or years-of-experience filter. Pass canonical slugs (jobFamilySlug, locationSlug, countrySlug, companySlug): resolve names first with search-job-families, search-locations and search-companies; never guess a slug. REQUIRED arguments: jobFamilySlug, countrySlug.
send-feedback
ClaudeReport wrong or missing Levels.fyi data, or a missing feature, to the Levels.fyi team. Call it only after telling the user what you will send and getting their agreement. Never include personal details; contactEmail only when the user wants a reply and gives it. Returns { received, id }. REQUIRED arguments: kind, message.
upload-resume
ClaudeAdd or replace the signed-in user's resume on Levels.fyi. Call this only when the user asks to upload their resume. Before you call, tell the user Levels.fyi uses it to fill in their profile and match them with jobs and recruiters. Never upload a resume for another person. Accepts a PDF (at most 5 pages) or a DOCX, at most 4.5 MB. Set fileType from the file's actual extension: the upload must be a file of that format, or it cannot be read. Returns two ways to upload: - upload: a presigned S3 upload that expires in 10 minutes. Use it only if you can read the user's file and make HTTP requests yourself, e.g. a terminal agent with the file on disk: send an HTTP POST with multipart/form-data to upload.url, with every entry of upload.fields as form fields, followed by the file in a field named "file". A 201 response means it worked; a larger file is refused. - webUploadUrl: otherwise, give the user this link to upload the file themselves after signing in. Never paste the resume's text or contents anywhere else. The resume is parsed in the background, usually within a couple of minutes. Call get-profile afterwards to check it; until parsing finishes, it can still show the previous resume. REQUIRED arguments: fileType.
