- S3-compatible endpoint (preferred) — ideal if your data team can push to an S3-compatible bucket using standard AWS SDKs or tools
- SFTP — for teams that require traditional SFTP file transfer
Option A: S3-Compatible Endpoint (Preferred)
Cora provides a dedicated S3-compatible bucket for your organization. Any tool or SDK that supports S3 (AWS CLI, Boto3, custom ETL pipelines, etc.) will work out of the box.What Cora Provides
Your Cora team will send the following credentials via a secure link:- Endpoint URL — the S3-compatible endpoint
- Bucket name — scoped to your organization
- Access Key ID
- Secret Access Key
- Region
What You Need to Do
1
Receive your credentials from the Cora team
2
Configure your data pipeline or script to push files to the provided endpoint
3
Push a test file and confirm with your Cora contact that it was received
4
Begin regular delivery on the agreed schedule
Option B: SFTP
Cora provides a dedicated, hosted SFTP endpoint for your organization. No infrastructure setup is required on your end; just connect and push files.What Cora Provides
- Host — a dedicated subdomain (for example,
yourcompany.sftp.cora.ai) - Port — 22
- Username — provided by Cora
- Authentication — SSH public key
What You Need to Do
1
Generate an SSH key pair (ECDSA, Ed25519, or RSA in OpenSSH format)
2
Send the public key to your Cora contact via a secure link
3
If your organization requires IP allowlisting, provide the static IP addresses that will connect
4
Once Cora confirms the key is uploaded, test connectivity and push a sample file
5
Confirm with your Cora contact that the file was received
6
Begin regular delivery on the agreed schedule
File Format Requirements
CSV, Wide Format
We expect CSV files in wide format — one row per entity (for example, account), with metrics as columns:account_id+snapshot_dateshould serve as a composite key for deduplication- New metrics can be added as new columns; our pipeline picks them up automatically
- Avoid per-metric directory structures (for example, one folder per metric); use columns instead
Nested or Relational Data
If your data includes detail records tied to accounts (for example, individual transactions, usage events, or line items), we support three approaches:- JSON column — for small, bounded nested data within the main CSV
- Separate detail feed — a second CSV file with
account_idas a foreign key, for high-volume or complex related data - Aggregated summary columns — pre-aggregated values on the account row, when raw detail isn’t needed
Directory Structure
Keep it simple:- One folder per use case or agent workflow (for example,
account_metrics/,renewal_data/) - Daily snapshot files named
snapshot_YYYYMMDD.csv(for example,snapshot_20260317.csv)
Data Schema
Before you build your export, your Cora implementation team will send a detailed field list tailored to your use case. This typically includes:- Identity fields — account ID, account name
- Eligibility / gating attributes — flags that determine which accounts qualify
- Timing fields — snapshot date, cadence anchors (for example, renewal date)
- Suppression flags — fields that indicate when outreach should be paused
- Contact fields — names and emails for the relevant stakeholders
- Metrics — the specific data points that power the agent workflow
Security
- All credentials are shared via secure, expiring links; never over email or chat
- S3 credentials are scoped to your organization’s bucket only
- SFTP uses SSH public key authentication (password login is disabled)
- IP allowlisting is available for SFTP on request
Onboarding Checklist
Here’s the typical flow from start to first delivery:1
Align on the data schema (field list) with Cora for your use case
2
Build the export based on the schema
3
Cora provisions the delivery endpoint (S3 or SFTP) and sends credentials
4
Push a test file
5
Cora confirms receipt and validates the file format
6
Begin regular delivery on the agreed schedule