01 · INGEST
Ingest
OAuth-backed connectors pull posts, comments, and engagement from 15 platforms on a schedule or in real time.
Connectors · Datasets
Social Data Pipeline
Social data in.
Decision-ready
data out.
7i7n ingests posts insights from 15 platforms, structures them with purpose-built engines, and delivers typed rows to the warehouse your data team already uses. An optional Intelligence layer of AI agents sits on top.
No credit card
OAuth in minutes
Live · pipeline/prod-01
Ingest
connectors
Process
engines
Deliver
storages
payload.jsonl
stream
{ "source": "tiktok", "type": "post", "engine": "media", "deliver_to": "snowflake://warehouse/social.posts" }
throughput 2.4k posts/min
p95 latency 340 ms
Processed across 7i7n
Posts ingested, structured, and delivered since launch.
0
posts processed
2,418/min now
p95 latency 340 ms
The pipeline
Three stages, one pipeline.
Everything between the platform API and your warehouse, managed end-to-end. Add engines or change delivery targets without rewriting glue code.
02 · PROCESS
Process
Three structuring engines — documents, media, and search & insights — turn raw API payloads into typed, queryable rows.
Document · Media · Search
03 · DELIVER
Deliver
Eight file formats and eight storage targets, written with idempotent upserts and a replayable manifest.
Exporters · Storage Targets
Optional · opt-in per agent
Intelligence layer
AI agents that read the pipeline's output and answer questions about your social data. The pipeline runs fine without them — turn on skills per workspace only where they pay for themselves.
Ingest
connectors
Process
emits rows
Deliver
writes result
reads rows · writes results
Intelligence
AI agents · pluggable skills
Skill
Engagement Insights
Skill
Content ideas
Skill
Marketing brief
Intelligence reads from Process output and writes results back through Deliver. Every run is metered and logged.
Outcomes
What you can do with 7i7n.
Benchmark share of voice across platforms
Unified post and engagement schema lets analysts compare reach, sentiment, and velocity across 15 networks in one query.
Learn more →Ship agent-written briefs to marketing
Chain the marketing-brief skill onto a weekly cohort. 7i7n writes the doc, you approve, it lands in the PM's inbox.
Learn more →Hydrate a feature store without glue code
Write directly to Snowflake, BigQuery, or S3 with typed, de-duplicated rows and a replayable manifest.
Learn more →Audit every agent call against its output
Each intelligence run writes prompt, model, tokens, and result to a credit ledger so finance and data trust the bill.
Learn more →Integrations
Multiple sources. Multiple destinations.
15 OAuth connectors in, 8 storage targets out. Add a connector or a target in minutes with the SDK.
Sources
MMeta
IInstagram
TTikTok
XTwitter / X
YYouTube
+ 10 more
Ingest
0/min
Process
0/min
Deliver
0/min
Destinations
SnowflakeSF
BigQueryBQ
S3S3
PostgresPG
RedshiftRS
+ 3 more
Exporters: JSON · JSONL · CSV · TSV · XLSX · Parquet · Avro · ORC
2.4k posts/min flowing
Built for teams
Admin controls your platform team will sign off on.
Multi-tenant orgs
Isolate workspaces per brand or client with their own credentials, quotas, and audit log.
API keys with scoped access
Grant a key one connector, one agent, or one destination. Rotate or revoke without redeploying.
Role-based admin
Owners, operators, and analysts have different views. SAML SSO and SCIM included on every workspace.
Pay-as-you-go credits with ledger
Every agent call, ingest, and export debits credits to a visible ledger. No monthly fee, no surprise invoices.
Pay-as-you-go · credits
Pay only for what the pipeline runs.
7i7n is metered in credits. Buy a top-up when you need it, spend it per request, watch the ledger. No monthly fee, no seats, nothing to cancel.
How credits are spent
Looks granular — works out cheaper than any monthly or yearly plan, because you pay only for the requests you actually make. Every request is priced in micro-dollars (µ$) from three independent dimensions: CPU time on our engines, AI tokens consumed by agents, and bytes written to your destination. 1 credit = $0.01 = 10,000 µ$.
Ingests
1*
µ$ per CPU-ms
A 10-second ingest run ≈ 10,000 µ$ ≈ 1 credit ≈ $0.01. Metered across all 15 connectors; replays and backfills bill the same CPU time.
AI tokens
4*
µ$ per token
A 2,000-token summary skill ≈ 8,000 µ$ ≈ 0.8 credits ≈ $0.008. Debited only when the agent actually runs — engagement summary, post ideas, marketing brief.
Exports
27*
µ$ per KB
A 10 MB Parquet export ≈ 276,480 µ$ ≈ 27.6 credits ≈ $0.28. JSON, CSV, XLSX, Parquet, warehouse (SF, BQ, RS) all bill the same.
* These rates are illustrative averages and may change at any time. Get a precise estimate for your workload inside the app.
Top up when you need it
One-time credit purchases. Auto-refill is optional and off by default.
Starter
One-time
$10
= 1,000 credits
Enough to summarize ~2.5M tokens or ingest ~1M posts.
- All 15 connectors
- All agent skills & engines
- Community support
Team
Best value
$50
= 5,500 credits
10% bonus credits. Typical for 3–5 analysts running daily pipelines.
- +10% bonus credits
- Scoped API keys & multi-org
- Email support, 90-day history
Scale
One-time
$500
= 62,500 credits
25% bonus credits. For teams running warehouse-scale backfills.
- +25% bonus credits
- Priority ingest & export lanes
- Dedicated Slack channel
Enterprise
Annual commit · SLAAnnual commits at volume pricing, a dedicated region, and a 99.95% SLA. Credits pooled across the org; overage settled quarterly at your contract rate. SSO + SCIM aren't here — they're included on every workspace.
Volume credit rates
Dedicated region
EU / US data residency
99.95% SLA
Named solutions engineer
No monthly fee
Credits never expire
Usage visible per request
SSO + SCIM included
Nothing to cancel
FAQ
Questions your legal team will ask.
Who owns the data?
You do. 7i7n is a data processor — your raw payloads, agent outputs, and exports are yours to export, delete, or portable-export at any time. We retain only the metadata required to operate the pipeline.
What OAuth scopes do you request?
Read-only scopes specific to profiles, posts, comments, and insights. Per-connector scope lists are published in the docs; you can opt each connector into narrower scopes when your platform policy requires it.
Where is data stored and for how long?
Raw payloads land in an encrypted object store in the region you pick. Default retention is 90 days and is configurable per workspace — or set it to zero-retention pass-through if you deliver straight to your warehouse.
EU or US hosting?
Both. Pick us-east-1 or eu-west-1 at workspace creation; Enterprise customers can pin additional regions. Cross-region transfer is opt-in.
How does pay-as-you-go billing work?
You buy credits once ($10, $50, or $500 to start) and spend them per request: 0.4 credits / 1k AI tokens, 1 credit / 1k ingested posts, 0.2 credits / GB exported. 1 credit = $0.01. There is no monthly fee, no seat fee, and no contract — credits never expire and unused credits are refundable within 30 days. Every debit is written to the usage ledger so you can audit any charge down to the request that caused it.
What are the rate limits?
Default 600 req/min per workspace, lifted on request for Enterprise. Ingest is bounded by each platform's own limits; we back off and resume automatically. Agent calls are metered per request, not per minute.
Is SSO available, and do I need Enterprise to enable it?
Yes — every workspace can enable SAML 2.0 SSO (Okta and Azure AD certified, multi-IdP supported) and SCIM 2.0 user provisioning. Enterprise is only required for annual commits, dedicated regions, and SLAs.
When do I need the Enterprise tier?
When you want annual commits at volume pricing, a dedicated pinned region, sustained high-volume throughput, or a signed 99.95% SLA. SSO and SCIM are not part of the Enterprise gate — they're available on every workspace.
Get started
Wire up your first pipeline
in under 10 minutes.
OAuth a connector, pick an agent, choose a destination. That's the whole setup.
No monthly fee · 100 free credits on signup · Pay only for what you run