Developer intelligence · GitHub · July 7, 2026
972,576 enriched GitHub developer profiles — followers, company, interests, recent activity and clean geo on 97.1% of located accounts — queryable over one REST API. Filter by country, stack, follower band or reachability; sort by influence; run a radius search. Public profile fields only, pay on success.
972,576
developer profiles — one record per login, enriched and scored.
26.4%
geo-resolved
93.6%
active 90d
4.2y
avg age
Snapshot July 7, 2026 — public fields only, suspected-automation hidden by default.
972,576
enriched GitHub developer profiles, deduplicated to one record per login and averaging 4.2 years old. A default query returns 970,617 — 1,959 suspected-automation accounts are hidden unless you ask for them.
97.1%
of accounts that give a location are cleanly reverse-geocoded to a country — 256,502 profiles carry a resolved geo.country, from the 264,037 (27.2%) that publish any location at all.
93.6%
pushed, opened a PR or reviewed in the last 90 days — this is a live-developer index, not a graveyard of dormant sign-ups. 21.2% expose a public contact channel.
311K
followers on the most-followed account (torvalds), but the median developer has a handful — 95.3% sit under 100 followers. Influence is a very long tail.
Of the 256,502 profiles with a resolved country, the United States leads (17.2%) with India close behind (12.8%) — then Brazil, China and Germany. GitHub's developer base is genuinely worldwide, and the geo backfill means you can target essentially any account that states a location.
| United States (US) | 43996 |
| India (IN) | 32915 |
| Brazil (BR) | 15069 |
| China (CN) | 13205 |
| Germany (DE) | 11861 |
| United Kingdom (GB) | 9561 |
| France (FR) | 8329 |
| Canada (CA) | 7311 |
| Indonesia (ID) | 5771 |
| Japan (JP) | 5684 |
| Spain (ES) | 5117 |
| South Korea (KR) | 5026 |
| Australia (AU) | 4080 |
| Vietnam (VN) | 3935 |
| Netherlands (NL) | 3655 |
| Bangladesh (BD) | 3579 |
| Italy (IT) | 3361 |
| Poland (PL) | 3313 |
| Russia (RU) | 3309 |
| Argentina (AR) | 3180 |
| Pakistan (PK) | 3066 |
| Switzerland (CH) | 2546 |
| Nigeria (NG) | 2414 |
| Colombia (CO) | 2146 |
| Sweden (SE) | 2126 |
| Mexico (MX) | 2120 |
| Türkiye (TR) | 2048 |
| Ukraine (UA) | 1870 |
| Taiwan (TW) | 1788 |
| Philippines (PH) | 1727 |
| Portugal (PT) | 1516 |
| Egypt (EG) | 1499 |
| Singapore (SG) | 1449 |
| Norway (NO) | 1431 |
| Sri Lanka (LK) | 1426 |
| Czechia (CZ) | 1355 |
| Belgium (BE) | 1294 |
| Kenya (KE) | 1274 |
| Austria (AT) | 1248 |
| Thailand (TH) | 1129 |
| South Africa (ZA) | 1123 |
| Denmark (DK) | 1102 |
| Chile (CL) | 1088 |
| Israel (IL) | 1071 |
| Finland (FI) | 1012 |
| Peru (PE) | 992 |
| New Zealand (NZ) | 965 |
| Ireland (IE) | 943 |
| Malaysia (MY) | 868 |
| Romania (RO) | 854 |
Top 20 countries by geo-resolved developers
A profile is only useful if the fields you need are populated. Here is how many of the queryable records carry each signal — geo is now the richest, thanks to a full geocoding backfill; direct contact channels are rarer, as you would expect from a public dataset.
Followers on GitHub are extraordinarily concentrated. The average developer has 19 followers; the 99th percentile is 277 and the 99.9th is 2,076 — only at the very top do you reach the 9 mega accounts led by torvalds at 310,536. 95.3% of developers sit under 100 followers, which is why rank_score (a blend of reach, output and recency) is often a better sort than raw followers.
| Tier (followers) | Developers | Share |
|---|---|---|
| Nano (< 100) | 924,809 | 95.3% |
| Micro (100–999) | 40,826 | 4.2% |
| Mid (1K–9.9K) | 4,679 | 0.5% |
| Macro (10K–99K) | 294 | 0% |
| Mega (100K+) | 9 | 0% |
Each developer is tagged with the areas their public activity leans toward, so you can slice the population by stack. Backend, ML/AI and frontend dominate; Web3 and game dev are the smallest of the tagged domains.
Company is self-declared and normalized (lowercased, @-handles stripped), so counts are a floor, not a census. The city rollup uses the geocoded location.
Top employers by self-declared company
Top cities by geocoded location
The nine mega accounts and the top of the macro tier, straight from a followers_desc query. This is exactly what the API returns — nothing hand-picked.
| Login | Name | Followers | Company / country |
|---|---|---|---|
torvalds | Linus Torvalds | 310,536 | Linux Foundation · United States |
claude | Claude | 124,709 | Anthropic |
ruanyf | Ruan YiFeng | 86,825 | China |
sindresorhus | Sindre Sorhus | 79,849 | Open-source maintainer |
bradtraversy | Brad Traversy | 76,062 | Traversy Media · United States |
JakeWharton | Jake Wharton | 68,380 | United States |
lucidrains | Phil Wang | 61,064 | ML / AI · United States |
steipete | Peter Steinberger | 52,078 | United Kingdom |
addyosmani | Addy Osmani | 50,700 | Google · United States |
geohot | George Hotz | 46,890 | comma.ai · United States |
One enriched record per developer. Grouped for readability; the API returns a flat object with a nested geo block.
Identity
loginidtypenameavatar_urlhtml_urlProfile & work
companycompany_normalizedbioblogemailtwitter_usernamehireableLocation & geo
location_rawgeo.countrygeo.country_codegeo.stategeo.citygeo.location (lat/lon)Contact & reach
social_accounts[]has_emailhas_twitterhas_blogsocial_countreachableAudience & influence
followersfollowingfollower_following_ratioinfluence_tierrank_scoreOutput & activity
public_repospublic_gistscreated_ataccount_age_yearslast_active_atactive_90dpushes_30dprs_30dreviews_30dInterests & flags
domains[]is_orgis_botis_suspected_automationcrawled_atThe search endpoint takes the filters below; combine any of them. Page with page and page_size (≤100 per page, and page × page_size ≤ 10,000).
Full-text & identity
Geography
Reach & intent
Audience & activity
Hygiene
Sort
relevancerank_score_descfollowers_descaccount_age_descaccount_age_ascdistance_asc (with lat/lon)Every query authenticates with an x-api-key header and reads the stored search index — there is no live crawl or proxy to manage, and you are billed pay on success: charged for results, not failed requests. Four endpoints cover it:
GET /datasets/github-users/search — filter, sort, page the population.GET /datasets/github-users/items/{login} — one enriched profile.GET /datasets/github-users/facets — counts for any dimension.GET /datasets/github-users/nearby — developers within a radius.The same four are exposed as MCP tools — datasets_github_users_search, datasets_github_users_facets, datasets_github_users_item and datasets_github_users_nearby — and covered by all six client SDKs, so an agent can call them directly.
Cite this
Crawlora (2026). GitHub Users Dataset. 972,576 enriched developer profiles; public fields only. https://crawlora.net/datasets/github-users.
Search with filters
# Reachable ML/AI developers in Germany, most influential first
curl "https://api.crawlora.net/api/v1/datasets/github-users/search?domain=ml-ai&country=Germany&reachable=true&sort=rank_score_desc" \
-H "x-api-key: $CRAWLORA_API_KEY"One profile by login
# One enriched profile by login
curl "https://api.crawlora.net/api/v1/datasets/github-users/items/torvalds" -H "x-api-key: $CRAWLORA_API_KEY"Facet the population
# Facet the population by country (or company, city, domain, influence_tier …)
curl "https://api.crawlora.net/api/v1/datasets/github-users/facets?facet=country" -H "x-api-key: $CRAWLORA_API_KEY"Nearby, by radius
# Developers within 25km of central London, 100+ followers
curl "https://api.crawlora.net/api/v1/datasets/github-users/nearby?lat=51.5074&lon=-0.1278&radius_m=25000&min_followers=100" \
-H "x-api-key: $CRAWLORA_API_KEY"The same figures behind the map and bars — plain and machine-readable for search engines and AI answer engines that cannot parse a chart.
| Country | Developers | Share of geo-resolved |
|---|---|---|
| United States (US) | 43,996 | 17.2% |
| India (IN) | 32,915 | 12.8% |
| Brazil (BR) | 15,069 | 5.9% |
| China (CN) | 13,205 | 5.1% |
| Germany (DE) | 11,861 | 4.6% |
| United Kingdom (GB) | 9,561 | 3.7% |
| France (FR) | 8,329 | 3.2% |
| Canada (CA) | 7,311 | 2.9% |
| Indonesia (ID) | 5,771 | 2.2% |
| Japan (JP) | 5,684 | 2.2% |
| Spain (ES) | 5,117 | 2% |
| South Korea (KR) | 5,026 | 2% |
| Australia (AU) | 4,080 | 1.6% |
| Vietnam (VN) | 3,935 | 1.5% |
| Netherlands (NL) | 3,655 | 1.4% |
| Bangladesh (BD) | 3,579 | 1.4% |
| Italy (IT) | 3,361 | 1.3% |
| Poland (PL) | 3,313 | 1.3% |
| Russia (RU) | 3,309 | 1.3% |
| Argentina (AR) | 3,180 | 1.2% |
| Pakistan (PK) | 3,066 | 1.2% |
| Switzerland (CH) | 2,546 | 1% |
| Nigeria (NG) | 2,414 | 0.9% |
| Colombia (CO) | 2,146 | 0.8% |
| Sweden (SE) | 2,126 | 0.8% |
| Mexico (MX) | 2,120 | 0.8% |
| Türkiye (TR) | 2,048 | 0.8% |
| Ukraine (UA) | 1,870 | 0.7% |
| Taiwan (TW) | 1,788 | 0.7% |
| Philippines (PH) | 1,727 | 0.7% |
| Portugal (PT) | 1,516 | 0.6% |
| Egypt (EG) | 1,499 | 0.6% |
| Singapore (SG) | 1,449 | 0.6% |
| Norway (NO) | 1,431 | 0.6% |
| Sri Lanka (LK) | 1,426 | 0.6% |
| Czechia (CZ) | 1,355 | 0.5% |
| Belgium (BE) | 1,294 | 0.5% |
| Kenya (KE) | 1,274 | 0.5% |
| Austria (AT) | 1,248 | 0.5% |
| Thailand (TH) | 1,129 | 0.4% |
| South Africa (ZA) | 1,123 | 0.4% |
| Denmark (DK) | 1,102 | 0.4% |
| Chile (CL) | 1,088 | 0.4% |
| Israel (IL) | 1,071 | 0.4% |
| Finland (FI) | 1,012 | 0.4% |
| Peru (PE) | 992 | 0.4% |
| New Zealand (NZ) | 965 | 0.4% |
| Ireland (IE) | 943 | 0.4% |
| Malaysia (MY) | 868 | 0.3% |
| Romania (RO) | 854 | 0.3% |
| Field | Developers | Share |
|---|---|---|
| Any location string | 264,037 | 27.2% |
| Resolved geo.country | 256,502 | 26.4% |
| Reachable (any contact) | 205,889 | 21.2% |
| Blog / website | 164,684 | 17% |
| Open to hire | 73,102 | 7.5% |
| Twitter / X handle | 54,883 | 5.7% |
| Public email | 33,843 | 3.5% |
| Interest domain | Developers | Share |
|---|---|---|
| Backend | 34,916 | 3.6% |
| ML / AI | 28,603 | 2.9% |
| Frontend | 25,254 | 2.6% |
| DevOps | 12,947 | 1.3% |
| Mobile | 7,728 | 0.8% |
| Security | 6,829 | 0.7% |
| Data | 5,149 | 0.5% |
| Web3 | 3,547 | 0.4% |
| Game dev | 2,139 | 0.2% |
| Company | Developers |
|---|---|
| Red Hat | 1,804 |
| Microsoft | 1,745 |
| 816 | |
| Amazon | 658 |
| NVIDIA | 560 |
| IBM | 537 |
| Intel | 279 |
| Elastic | 249 |
| SAP | 246 |
| Adobe | 232 |
| City | Developers |
|---|---|
| Bengaluru | 3,792 |
| London | 3,563 |
| New York | 3,377 |
| San Francisco | 3,273 |
| Seoul | 2,778 |
| Berlin | 2,185 |
| Paris | 2,144 |
| Seattle | 1,924 |
| Dhaka | 1,915 |
| Beijing | 1,906 |
Query 973K developer profiles — followers, stack, activity and clean geo — through one REST API. Recruiting, developer-relations targeting, ecosystem research or lead lists: clean JSON, public fields only, pay on success.
The dataset holds 972,576 enriched developer profiles (schema v3), deduplicated to one record per GitHub login. A default search returns 970,617 of them — the other 1,959 are suspected-automation (commit-farm / mass-repo) accounts, hidden unless you pass is_suspected_automation=true. Organization and bot accounts are excluded from the crawl.
Every field is a public, credential-free profile field: login, name, company, bio, blog, avatar, followers/following, public repo and gist counts, account age and recent activity (pushes, PRs and reviews in the last 30 days), interest domains, an influence tier and a composite rank score, plus a reverse-geocoded location. Nothing behind authentication is collected. The email field is present only when the developer chose to publish an email on their own public profile — it appears on about 3.5% of records.
About 27.2% of profiles (264,037) publish a free-text location, and a geocoding pass with fuzzy-match guards resolves 97.1% of those to a real country — 256,502 profiles with a clean geo.country, plus state, city and coordinates where available. That means you can filter or run radius queries on essentially every account that says where it is, with far less noise than parsing the raw location string yourself.
influence_tier buckets each developer by follower count: nano (under 100), micro (100–999), mid (1K–9.9K), macro (10K–99K) and mega (100K+). It is a very long tail — 95.3% of developers are nano and only 9 accounts are mega. rank_score is a separate composite (followers, repos, recency) you can sort or threshold on for a blend of reach and activity.
Yes. The dataset is queryable over Crawlora's REST API and via MCP: full-text and faceted search, one enriched profile by login, facet counts for any dimension, and a geo-radius nearby query. Filter by location, company, interest domain, follower band, activity, reachability and more; sort by relevance, rank score, followers, account age or distance. All billed pay-on-success.
Profiles are continuously enriched; this snapshot is dated July 7, 2026. Each record carries its own crawled_at timestamp and a last_active_at, so you can tell how recently a given profile was refreshed and when the developer last pushed.