Open dataset

The startup credits database, free to download

Every startup credit, perk, and software deal we track, as a clean, documented database you can query, analyze, or build on. Curated from each provider's own official page, dated, confidence-tiered, and refreshed weekly. Available as JSON and CSV under CC BY 4.0: use it anywhere, just credit the source.

2,887 programs2,441 providers390 with a dollar value9 categories

Also on Hugging Face, Kaggle, and as a citable snapshot on Zenodo (DOI 10.5281/zenodo.23493683).

Updated daily · checked 36 min ago
What you get

One row per program, everything sourced

Each record carries the provider, category, headline benefit value, and the real eligibility gates (stage, funding cap, geography), plus a link straight to the provider's own page. Every value is the same figure the site shows, drawn from the same source, so the dataset never disagrees with the pages. For AI training jobs and pay, the team behind StartupPerks publishes the same kind of data on AITraining.jobs, built from an hourly read of the platforms' own job boards and sent as files by email.

Complete

All 2,887 tracked programs, not a sample. Filter to one category with a query parameter.

Sourced

Every record carries the address of the provider's official page, a confidence tier and the date it was last re-verified.

Free to use

Licensed CC BY 4.0. Use it in a product, a study, or a model, with attribution.

Schema

Every field in a record

The JSON returns these fields per record inside a documented envelope (license, attribution, methodology, counts). The CSV has one column per field, in this order; list fields are joined with a semicolon.

key
string

Stable unique identifier for the program.

name
string

Program name as the provider presents it.

provider
string

The company offering the perk.

provider_slug
string

Provider identifier (matches /providers/[slug]).

category
string

Category key, e.g. cloud-infra, ai-ml, banking-fintech.

category_label
string

Human-readable category label.

benefit_type
string

Kind of benefit: Credits, Discount, Free, Cash & rewards, etc.

application_status
string

open, paused, closed, or unknown. Based on an explicit source notice; unknown includes legacy records whose page was readable but application availability was not verified.

application_status_source_url
string | null

Official page supporting the application status, independently of the benefit-value source. Null when status is unknown.

application_status_checked_at
string | null

ISO timestamp when the application status was checked. Null when status is unknown.

offer_ends_on
string | null

Last day (YYYY-MM-DD) the offer or its current application window is open, as the provider's page states it. Once that day is over, application_status is closed. Null when the page states no end date.

value_text
string | null

Headline benefit value exactly as the provider states it.

value_usd
number | null

USD amount explicitly stated in value_text. Null for unpriced credit/token counts, inferred or calculated equivalents, non-USD amounts, partner-bundle totals, billing thresholds, and tax credits a service helps claim (the government pays those).

eligibility
string

Primary funding/backing gate, e.g. Any startup, VC / accelerator-backed.

stages
string[]

Startup stages that qualify (empty means any stage).

max_funding_raised_usd
number | null

Maximum total funding raised the program allows, if any.

max_company_age_years
number | null

Maximum company age in years the program allows, if any.

geographies
string[]

Regions the program is open to (empty means global).

recurring_or_expiry
string | null

Whether the benefit recurs or when it expires, as stated.

other_requirements
string | null

Any other stated requirement (e.g. new customers only).

confidence
string

How well-verified the record is: high, medium, or low.

last_verified
string | null

ISO timestamp this record's terms were last checked on the provider's own pages (by the freshness reader or by research), or null if never checked at the source.

terms_reads
number

How many times StartupPerks' reader has read the provider's page for this program since 2026-08-14 (every read is kept with its date).

terms_last_changed_on
string | null

Day (YYYY-MM-DD) a read last found the terms changed: a value figure the page stopped stating, or the program stopping or resuming applications. Rewording is not a change. Null when no read has found a change (see terms_reads).

source_url
string | null

The provider's official page stating the terms (the source of record).

apply_url
string | null

The provider's own application URL; for a program that accepted StartupPerks as a partner, the application link the provider issued to StartupPerks, which grants its partner tier. Null while applications are known to be paused or closed.

same_owner
string | null

Set when the program comes from a company run by the same company as StartupPerks, saying so in plain words; null for every other program. Such a program is listed free and ranked by the same rules.

url
string

The StartupPerks page for this program.

Use it

A sample record, and how to fetch it

Endpoints
# Full dataset as JSON
curl https://startupperks.co/api/dataset

# One category as CSV
curl "https://startupperks.co/api/dataset?format=csv&category=ai-ml"

# In JavaScript
const res = await fetch("https://startupperks.co/api/dataset");
const { records } = await res.json();
Sample record
{
  "key": "google-for-startups-cloud-program",
  "name": "Google for Startups Cloud Program",
  "provider": "Google Cloud",
  "provider_slug": "google-cloud",
  "category": "cloud-infra",
  "category_label": "Cloud & Infrastructure",
  "benefit_type": "Credits",
  "application_status": "unknown",
  "application_status_source_url": null,
  "application_status_checked_at": null,
  "offer_ends_on": null,
  "value_text": "Up to $350K credits",
  "value_usd": 350000,
  "eligibility": "Any startup",
  "stages": [
    "idea",
    "pre-seed",
    "seed",
    "series-a"
  ],
  "max_funding_raised_usd": null,
  "max_company_age_years": 5,
  "geographies": [],
  "recurring_or_expiry": "Scale credits span two years: 100% of usage up to the year-one amount, then a 20% rebate on year-two spend up to $100,000. Each credit expires on the date Google states in writing when it issues it, or the date shown in your console. One-time enrollment.",
  "other_requirements": "Start tier: founded in the past 2 years, with a working MVP, a clear business model and plans to seek venture funding soon. Funded tiers (Scale and Scale AI): verifiable Pre-Seed to Series A funding from a known investor, founded within the last 3 to 5 years depending on the tier, a working MVP, a clear roadmap and a verified company domain, and no more than $5,000 in earlier Google Cloud credits; angel, friends and family, grants, crowdfunding and private equity do not qualify for Scale. The AI tier: founded within the last 5 years, Pre-Seed/Seed rounds within the last 5 years or a Series A within the last 12 months (SAFEs qualify), and an AI-first product built on Gemini. Credits cover Google's own models, including the Gemini API, but not third-party models. No equity taken.",
  "confidence": "high",
  "last_verified": "2026-11-07T00:00:00.000Z",
  "terms_reads": 3,
  "terms_last_changed_on": null,
  "source_url": "https://cloud.google.com/startup",
  "apply_url": "https://cloud.google.com/startup",
  "same_owner": null,
  "url": "https://startupperks.co/programs/google-for-startups-cloud-program"
}
What's inside

Programs by category

The dataset spans every category on the site. Add ?category=<key> to either endpoint to pull just one.

License

Free to use, with credit

The dataset is published under Creative Commons Attribution 4.0. You can use it commercially, remix it, and build products on it. The one requirement is attribution: credit StartupPerks and link back.

CC BY 4.0Attribution required

Suggested credit line, copy as-is:

Data from StartupPerks (https://startupperks.co), licensed CC BY 4.0.

Citing it in a paper or report? Cite the fixed snapshot with its DOI:

StartupPerks (2026). StartupPerks Open Dataset: Startup Credits, Perks and Deals (2026-10-20 snapshot) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23493683
Figures are the headline maximums each provider advertises and can change; the dataset records when each was last checked and the address of the provider's own page. See the methodology for how every value and eligibility rule is sourced.