Web data infrastructure

Cast a net
across the
open web.

Seine pulls real-time, structured data from any website, API, or source — and lands it in your stack clean and query-ready. One API for the whole catch.

Landing now · live
0 records
across {{12,000+}} sources · p95 latency {{280}}ms · since you opened this page
source: example.com
caught 1,204 records
schema: normalized ✓
dupes dropped: 312
Dive in
Coverage

One net. Any source.

If it's on a public page, Seine can land it — normalized to a single schema you never have to babysit.

{{E-commerce}} {{Marketplaces}} {{Social}} {{News & media}} {{Search / SERP}} {{Directories}} {{Job boards}} {{Real estate}} {{Reviews}} + anything public
0
Records landed / day
{{280}}ms
p95 API latency
0
Field completeness
{{12k}}
Sources supported
The platform

Three ways to bring the data in.

01 / API

Real-time API

Query any page, listing, or endpoint the moment it changes. REST + GraphQL, cursor-paginated, rate-fair. Sub-second on the hot path.

RESTGraphQLOAuth
02 / BULK

Datasets

Historical and bulk. Backfill years of public pages as Parquet, JSONL, or a direct sync into your warehouse. Refreshes on your cadence.

ParquetSnowflakeS3
03 / STREAM

Live streams

Subscribe to a site, search, or feed. We push structured events the second a page changes — no polling, no missed beats. Webhooks + Kafka.

WebhooksKafkaSSE
How it works

Cast. Filter. Land.

A seine doesn't snare one fish at a time — it surrounds the whole shoal. Same idea, for data.

// 01
Cast

Spread wide, go deep

Point Seine at a site, a search, or a list of URLs. We fan out across the surface and pull every public record in range.

// 02
Filter

The noise stays out

Noise, boilerplate, and duplicates never reach you. Every record is validated, de-duplicated, and mapped to one consistent schema.

// 03
Land

Clean, typed, ready

The catch arrives in your warehouse, API, or stream — typed and query-ready in {{seconds}}, not pipelines you maintain.

The catch

One schema. Any source.

GET /v2/posts/{id} · 200 OK · {{41}}ms
// every record, the same shape — whatever the source
{
  "id":          "sn_3f9a…c2",
  "source":      "{{example.com}}",
  "url":         "https://{{example.com}}/p/8421",
  "type":        "product",
  "captured_at": "2026-06-08T14:22:07Z",
  "title":       "{{ extracted title … }}",
  "price": {
    "amount":   129.00,
    "currency": "USD"
  },
  "in_stock":    true,
  "lang":        "en",
  "text":        "{{ normalized body … }}",
  "entities":    ["{{category}}", "{{brand}}"]
}
  • Normalized, not rawThe same fields whether it came from a marketplace or a news site. Map once.
  • Enriched on arrivalLanguage, entities, and key fields parsed before the record ever reaches you.
  • Stable IDs, no dupesDeterministic IDs across runs. Stream it for a year — never store the same page twice.
  • SDKs that fit your stack{{Python}}, {{TypeScript}}, {{Go}}, and a CLI. Typed clients, generated from the spec.
Who casts with us

Signal for teams that move first.

E-commerce · Pricing

Watch every price

Track competitor catalogs, stock, and pricing across thousands of stores in real time.

Finance · Alt-data

Trade on what's public

Turn web signals — prices, listings, filings — into tradable data before the market reprices.

Research · Insights

Map the market

See what a whole market is doing — at population scale, not survey scale.

AI · Model teams

Feed your models

Fresh, structured, deduplicated corpora — the raw material for training and eval.

Growth · Lead gen

Build better lists

Pull firmographics, contacts, and intent from public sources — deduped and verified.

SEO · SERP

Track the rankings

Monitor search results, snippets, and competitors across regions and devices.

Responsible by design

Built to be trusted with data.

Scraping public web data is a serious responsibility. Seine is engineered to keep you on the right side of it — by default, not as an afterthought.

{{Draft copy — confirm every claim with legal before publishing. Compliance posture varies by jurisdiction, platform terms, and use case.}}

  • Public data only. No logins, no private content, no walled gardens.
  • GDPR & CCPA aligned. Lawful-basis tooling, data-subject requests, regional storage.
  • Built-in opt-out & takedown. Honoured across every dataset and stream, automatically.
  • PII minimization. Redact-on-request and field-level controls baked into the API.
  • {{SOC 2 Type II}} · audit logs. Every request traceable, every access accounted for.
Pricing

Pay for the catch, not the boat.

Usage-based, transparent, and built to grow from a side project to a fleet.

Cast
For builders testing the waters.
{{$0}} / forever
  • {{5,000}} records / month
  • Real-time API + bulk export
  • Community support
  • One API key
Haul Popular
For teams shipping on social data.
{{$499}} / month
  • {{2,000,000}} records / month
  • Any source + live streams
  • Webhooks, Kafka & warehouse sync
  • {{99.9%}} uptime SLA
  • Priority support
Fleet
For scale, compliance & volume.
Custom
  • Unlimited records
  • Dedicated capacity & on-prem option
  • Custom sources & backfills
  • DPA, security review, SSO
  • Named solutions engineer

Ready to bring in the catch?

Spin up a key, run your first query in under five minutes, and see what the social web has been saying all along.