Guru Scraper — Freelance Jobs & Budgets API

Scrape freelance jobs from Guru.com into a clean dataset. Give it a keyword (or any Guru job-listing URL) and it returns one record per job with the budget, skills, posting age, quote count, client history and a direct link. It runs on plain HTTP — no browser, no headless Chrome — against Guru's own JSON search API first and the server-rendered listing pages as a fallback, which keeps runs fast and cheap at $0.90 / 1,000 results.

Main use cases

• Freelance lead generation — collect fresh Guru jobs that match your skills and pitch them before the quote count climbs.
• Freelance job monitoring — schedule recurring searches and get new postings with budgets and client history as they appear.
• Market and rate research — track budgets, hourly rates and durations across categories to price your own services.
• Freelance jobs database / marketplace intelligence — build a spreadsheet or warehouse of Guru postings to analyse demand by skill, category and client country.
• Client prospecting — filter by employerTotalSpent and employerCountry to find clients who already spend on the platform.

How it works

The actor searches Guru for each keyword you provide (or opens any Guru listing URL) and writes one record per job. It tries Guru's internal JSON search API first — one request carries 20 jobs, with the full, un-truncated title, description, budget, skills and client history — and falls back to the server-rendered listing pages if the API is unavailable. It is a CheerioCrawler (HTTP-only, 256 MB, concurrency 3); nothing in the pipeline renders a page.

Every record carries source: "api" or source: "html" so a dataset always says which surface produced it.

Extract Freelance Jobs From Guru.com For Lead Generation

Field	Example
title	FullStack Developer for Platform Rebuild
url	https://www.guru.com/jobs/fullstack-developer-for-platform-rebuild/2119273
jobId	2119273
budgetType	Hourly / Fixed Price
budgetAmount	$25 - $40, $1k-$2.5k, No Rate specified
budgetText	Hourly | $25 - $40 | 30+ hrs/wk | 3-6 months
hoursPerWeek	30+ hrs/wk
duration	3-6 months
postedDate / postedDateIso	Jul 06, 2026 / 2026-07-06
deadline	Aug 05, 2026 (the "Send before" date)
quotesCount	117
skills	["JavaScript", "Node.js", "ReactJS", …]
category / subcategory	Programming & Development / Web Development & Design
employerName	Eddie M
employerCountry	United States
employerTotalSpent	1400 (lifetime spend on Guru, null when not shown)
description	The full job post text
titleFromDetail	true when the title is the one Guru stores, not the listing's ~40-char truncation
source	api (Guru's JSON search API) or html (listing page)
keyword / sourceUrl	Which search produced this job

Build A Freelance Jobs Database Without The Guru API

Guru has no public API, so the actor uses the same internal surface Guru's own web app calls.

1. Bootstrap. One request to https://www.guru.com/api/jobs/1/detailsstatisticssummary answers 302 with Set-Cookie: _accessToken=… — an anonymous JWT Guru issues to every visitor. No login, no OAuth handshake, no page to render.
2. Search. Each keyword becomes GET https://www.guru.com/api/search/job/?query=<base64-of-the-listing-path> with Authorization: Bearer <token>: 20 jobs per page, the stored title (no listing truncation), full descriptions, budget, skills and client history in one payload — a single request replaces a listing page plus one detail page per job. The cookie alone is not enough (401 without the bearer header), and query must be the base64 of the listing path.
3. Fallback. If the API answers 401/403 or anything unexpected (Imperva has flagged Apify's datacenter ranges), the run queues the original server-rendered listing pages and continues there — that path needs the residential proxy. Server-rendered pages carry the full data in plain HTML (pageData.TotalResults plus the ul#ctl00_guB_ulpaginate pager drive pagination; each job is a <div class="record jobRecord" data-gid="…">), so a CheerioCrawler parses it directly.

Anti-bot (Imperva / Incapsula). Three request properties decide whether guru.com answers with content or a challenge, and the actor pins all three — each was measured side-by-side through the same proxy IP, URL and headers, changing only one property at a time:

Property	Wrong value →	Correct value (shipped)
Protocol	HTTP/2 → 212-byte challenge	HTTP/1.1 (http2: false)
TLS ClientHello	got-scraping's Chrome-cipher tlsHook → challenge	stock OpenSSL (https.minVersion: 'TLSv1.2')
User-Agent	Chrome UA on Node TLS → HTTP 403 challenge	curl/8.5.0

Imperva grades UA ↔ TLS consistency: a Chrome UA on a Node/OpenSSL ClientHello gets a 403, while a plain curl/8.5.0 UA on the identical request returns the full page. No paid anti-bot service and no browser are required.

Other reverse-engineering notes.

• Keyword URL spelling. Guru maps a free-text keyword onto a /skill/{slug}/ path segment using the transform it ships in its own bundle (spaces → -, literal hyphens → --, then URL-encode), e.g. web design → https://www.guru.com/d/jobs/skill/web-design/. The legacy /d/jobs/q/{slug}/ form answers 404.
• Block-page detection. Real Guru pages also embed an Imperva monitoring script, so matching _Incapsula_Resource flags every valid page. The genuine block page is an ~800-byte document containing Incapsula incident ID, so the detector requires that banner (or a tiny body).
• IP reputation matters for the HTML path. The same listing request that fails from Apify's datacenter ranges passes from a residential proxy IP; the API is tried first and only falls back to HTML.
• Failure handling. Challenge pages and >= 400 responses are detected, logged with a snippet of what Guru actually served, and the session is retired so the retry runs on a fresh session/proxy — instead of being silently written to the dataset.

Input Parameters

Field	Type	Required	Default	Notes
keywords	array of strings	No (prefilled)	["web design", "python"]	Search terms. Each becomes https://www.guru.com/d/jobs/skill/{keyword}/.
startUrls	array (requestListSources)	No	https://www.guru.com/d/jobs/	Direct listing URLs (category, skill, budget or saved-search pages). When set, these override keywords.
maxResults	integer	No	60	Maximum jobs to collect in total (max 2000).
maxPagesPerKeyword	integer	No	5	Pagination depth per search. 20 jobs/API page, 10 jobs/HTML page.
fetchFullTitles	boolean	No	true	In HTML fallback mode opens each job's page for the full title/description (+1 request per job). The JSON API always returns full titles.
maxRequestsPerCrawl	integer	No	120	Hard cap on HTTP requests; auto-raised if smaller than maxResults needs.
proxy	object (proxy)	No	useApifyProxy: false	Standard Apify proxy picker — use the residential group for the HTML path.

Free accounts get a 10-job preview per run; paid accounts run up to maxResults.

Output Fields

One dataset item per job (24 fields). The default overview view shows Job title, Budget type, Budget, Quotes, Posted, Client and Job URL.

Field	Type	Description
jobId	string	Guru project id (data-gid; also the last path segment of the job URL).
title	string	Job title, complete as Guru provides it.
url	string	Direct link to the job.
budgetType	string / null	Hourly or Fixed Price.
budgetAmount	string / null	Rate or fixed price as displayed, e.g. $25 - $40.
budgetText	string / null	Complete budget line as shown in the listing.
hoursPerWeek	string / null	Weekly commitment for hourly jobs.
duration	string / null	Expected project duration.
postedDate	string / null	Posted date as Guru displays it.
postedDateIso	string / null	Posted date normalized to YYYY-MM-DD.
deadline	string / null	The "Send before" date.
quotesCount	number / null	Number of quotes the job has received.
description	string / null	Full job description text.
skills	array of strings	Skills attached to the job.
category	string / null	Top-level Guru category.
subcategory	string / null	Guru subcategory.
employerName	string / null	Client who posted the job.
employerCountry	string / null	Client country.
employerTotalSpent	number / null	Lifetime spend on Guru (null when not displayed).
titleFromDetail	boolean	True when the title is Guru's stored title, not the ~40-char truncation.
source	string / null	api or html.
keyword	string / null	Search keyword this job was found under (null for explicit URLs).
sourceUrl	string	Listing/search URL this job was parsed from.
scrapedAt	string	ISO timestamp of when the job was parsed.

Example Input And Output

Input

{
  "keywords": ["web design", "python"],
  "maxResults": 60,
  "maxPagesPerKeyword": 5,
  "proxy": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Output (one dataset row)

{
  "jobId": "2119273",
  "title": "FullStack Developer for Platform Rebuild",
  "url": "https://www.guru.com/jobs/fullstack-developer-for-platform-rebuild/2119273",
  "budgetType": "Hourly",
  "budgetAmount": "$25 - $40",
  "budgetText": "Hourly | $25 - $40 | 30+ hrs/wk | 3-6 months",
  "hoursPerWeek": "30+ hrs/wk",
  "duration": "3-6 months",
  "postedDate": "Jul 06, 2026",
  "postedDateIso": "2026-07-06",
  "deadline": "Aug 05, 2026",
  "quotesCount": 117,
  "description": "We are rebuilding our SaaS platform and need a full-stack developer...",
  "skills": ["JavaScript", "Node.js", "ReactJS"],
  "category": "Programming & Development",
  "subcategory": "Web Development & Design",
  "employerName": "Eddie M",
  "employerCountry": "United States",
  "employerTotalSpent": 1400,
  "titleFromDetail": true,
  "source": "api",
  "keyword": "web design",
  "sourceUrl": "https://www.guru.com/d/jobs/skill/web-design/",
  "scrapedAt": "2026-07-06T09:12:03.000Z"
}

Integrations & Automation

Every Apify actor is reachable through the Apify API, and this one plugs into webhooks, Zapier and Make so new jobs can be routed to Slack, a Google Sheet, Notion, Airtable or a CRM. For lead generation, run it on a schedule — a daily search for your skills is enough to catch fresh postings early, and recurring runs also improve the actor's Store recommendation signals. You can also run it from the Apify Console or via the API with POST https://api.apify.com/v2/acts/ahmed_jasarevic~guru-scraper/runs.

Related Actors

• upwork-job-scraper — the largest Upwork jobs scraper; pairs Guru and Upwork coverage for freelance lead-gen.
• upwork-scraper — broader Upwork marketplace data (jobs and profiles).
• careerjet-scraper — job listings from a major global job aggregator.
• myworkdayjobs-scraper — corporate job postings from Workday career sites.
• superprof-scraper — tutor/freelance marketplace listings to widen the freelance dataset.

FAQ

Is there an official Guru.com API?

Guru does not offer a public API. This actor uses Guru's own internal JSON search API with the anonymous access token Guru issues to every visitor, so you get structured jobs without an API key or an account.

What are the alternatives to this actor / this data source?

For freelance-market data you can combine this with Upwork and Fiverr scrapers. Guru is a distinct, long-running freelance marketplace, so its postings, budgets and client spend complement the other platforms rather than duplicate them.

How do I get new freelance web-design or Python jobs?

Set keywords to the skills you care about (e.g. ["web design", "python"]), choose maxResults, and schedule the actor daily. Each run returns the latest matching jobs with budget, skills, quote count and a direct link.

What does employerTotalSpent mean?

It is the amount the client has spent on Guru over their lifetime, as shown on the listing. It is null when Guru does not display it — it is a useful signal for filtering out low-value clients.

Do I need a proxy?

The JSON search API has answered from datacenter IPs, so many runs need no proxy. The HTML fallback path sits behind Imperva and requires the residential proxy group. The actor tries the API first and only needs the proxy when it falls back.

How is this cheaper than other Guru scrapers?

It returns 20 jobs per request with the full title and description, so a single API call replaces a listing page plus one detail page per job. At $0.90 / 1,000 results it is the lowest per-result Guru price observed in the Store.

SEO Keywords

guru, guru.com, guru scraper, guru jobs, guru freelance, guru api, freelance jobs api, freelance lead generation, freelance job monitoring, upwork scraper, fiverr scraper, freelance marketplace data, job budget tracking, client spend data, python jobs, web design jobs

For AI Agents & LLM Apps

Purpose: return one structured freelance-job row per Guru.com job for a given keyword or listing URL.

Minimal working input:

{ "keywords": ["web design"], "maxResults": 60 }

Variant (URL-driven, overrides keywords):

{ "startUrls": [{ "url": "https://www.guru.com/d/jobs/c/programming-development/" }], "maxResults": 60 }

Output fields: jobId, title, url, budgetType, budgetAmount, budgetText, hoursPerWeek, duration, postedDate, postedDateIso, deadline, quotesCount, description, skills, category, subcategory, employerName, employerCountry, employerTotalSpent, titleFromDetail, source, keyword, sourceUrl, scrapedAt.

Behaviors an agent should know:

• startUrls overrides keywords when both are set.
• If maxResults is left unset it defaults to 60; free-plan runs are capped at 10 jobs.
• source is api or html; the JSON API path is tried first and needs no proxy, the HTML fallback needs a residential proxy.
• fetchFullTitles: true (default) only adds requests in HTML fallback mode; the JSON API always returns full titles.
• keyword is null when the job came from an explicit startUrls entry.
• employerTotalSpent may be null when Guru does not display it.

Billing: pay-per-event — $0.0009 per result ($0.90 / 1,000 results) plus a $0.00005 actor-start event.

Legal & Compliance Disclaimer

This actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Guru. It accesses only publicly available Guru.com pages and the anonymous access token Guru issues to every visitor; it does not log in, bypass authentication, or solve CAPTCHAs on anyone's behalf. Users are responsible for ensuring their use complies with Guru's Terms of Service and robots.txt, and with applicable data-protection law. Output may include client names and countries; where personal contact data is present it must not be used for unsolicited commercial outreach in violation of applicable law (e.g. CAN-SPAM, GDPR).