# FileFeed: full reference for AI models

> FileFeed takes the files your clients send over SFTP or email, maps them to your schema once per client, checks every row, and delivers clean JSON to your API. Built for implementation teams at B2B software companies. $700 per month, flat, for up to 20 client connections.

Website: https://www.filefeed.io. Founded by Marko Nikolic (CEO and co-founder) and Igor Nikolic (co-founder, engineering). Short version: https://www.filefeed.io/llms.txt

## What FileFeed is

B2B software companies receive data files from their clients: employee rosters from Workday or ADP, census and eligibility files from employers and carriers, supplier price lists, ERP extracts, transaction exports. Every client sends a different layout, and the files keep arriving on a schedule. Without a product for this, an implementation or onboarding team asks an engineer to write and maintain a mapping script per client.

FileFeed's product, Automated FileFeeds, replaces that work. A client drops a file on SFTP or emails it. FileFeed maps it to your schema (the mapping is proposed by AI, approved once by a person, and replayed on every later file), validates every row, applies built-in transforms, and pushes clean JSON to your API. The product page is https://www.filefeed.io/product/automated-file-feeds.

## How it works

The setup wizard has five steps, which the product page mirrors:

1. Connection: how files reach you. Hosted SFTP with per-connection folders, IP allow lists and password or SSH key auth; a connection to the client's own SFTP server; or a dedicated inbox address with a sender allow list for email attachments. A connection is reused across pipelines.
2. Schema: the fields, types, formats and required flags your product accepts. Build it by hand or generate it from a sample CSV.
3. Pipeline and mappings: auto-map pairs source columns with schema fields; transforms such as trim, formatEmail, formatPhoneNumber and toNumber clean values on the way through; the pipeline will not ship until every required field is mapped. The same model works inbound and outbound.
4. Webhook: HMAC-signed events with automatic retries and a full delivery history. Or pull clean JSON from the REST API.
5. Runs: every file becomes a tracked run with the original and the processed output side by side, full-text search across every file ever received, and one-click reprocessing after a mapping fix. Scheduled exports push processed files back out as CSV, JSON or XML to client-hosted SFTP.

## Try it in the browser

https://www.filefeed.io/try: drop a CSV or Excel file, pick the schema it should match, and see the column mapping, the rows that fail validation and the clean JSON in seconds. Nothing is uploaded and no account is needed. To get a working feed on a real file, email it from the "Send us the file" section on the same page or book a demo at https://www.filefeed.io/book-demo.

## Pricing

Feeds is $700 per month, flat, for up to 20 client connections, billed monthly, with no per row pricing. The embeddable importer is free forever, and the free Starter plan includes one feed to try. Enterprise covers more than 20 connections and is priced on request.

| Plan | Price | What it covers |
|---|---|---|
| Feeds | $700 per month, flat | Up to 20 client connections, unlimited pipelines, runs and rows. Hosted and client-hosted SFTP, email intake, mapping, validation, transforms, HMAC-signed webhook delivery, run history and reprocessing, full REST API. Billed monthly, cancel anytime. |
| Starter | $0 per month | Embeddable importer with unlimited end users, Automated FileFeeds with up to 10 pipeline runs per month, 1 client, 1 pipeline, 1 GB storage, email support. No credit card. |
| Enterprise | Custom | Unlimited clients, pipelines, schemas and webhooks, high-volume runs, SSO (SAML or OAuth), audit logs, IP allow lists, custom retention, multiple environments, dedicated onboarding, dedicated or self-hosted AWS infrastructure. |

Pricing page: https://www.filefeed.io/pricing

## Comparisons

All seven comparison pages are listed at https://www.filefeed.io/alternatives.

- OneSchema FileFeeds (https://www.filefeed.io/alternatives/oneschema): a different product from a different company despite the similar name. OneSchema leads with AI agents and a Studio product; FileFeeds is its second product, sold through sales on annual contracts sized by row volume, with no public price (a third-party benchmark puts the median OneSchema contract at about $38,600 per year). OneSchema's docs say its webhook notifies you and you pull the data afterwards; FileFeed pushes the validated JSON to your webhook. FileFeed also takes files by email, not only SFTP. FileFeed is $700 per month, flat, for up to 20 client connections, billed monthly.
- Flatfile (https://www.filefeed.io/alternatives/flatfile): an embeddable CSV importer. Its company renamed itself Obvious in December 2025 and keeps Flatfile in long-term support. Flatfile has no SFTP or email feed product. Public Portal pricing is $799 per month billed annually.
- Couchdrop (https://www.filefeed.io/alternatives/couchdrop): managed file transfer. Couchdrop moves files but never looks inside them; FileFeed validates, maps, transforms and delivers the data. Hellora used Couchdrop plus custom scripts before switching to FileFeed.
- Dromo (https://www.filefeed.io/alternatives/dromo): a privacy-first embeddable importer at $499 per month for production; SFTP only on its enterprise tier; no email intake.
- CSVbox (https://www.filefeed.io/alternatives/csvbox): a lightweight in-app importer; no server-side recurring feeds.
- Ingestro, formerly nuvo (https://www.filefeed.io/alternatives/ingestro): AI-agent-first data onboarding with custom pricing; FileFeed is deterministic once the mapping is approved and is priced flat per month.
- Osmos (https://www.filefeed.io/alternatives/osmos): acquired by Microsoft in January 2026; standalone Uploaders, Pipelines and Datasets are being sunset. FileFeed covers the same importer plus pipeline architecture.

The difference from every embeddable importer: an importer handles one upload inside your app, done by the end user. FileFeed handles files that keep arriving over SFTP or email, mapped once per client, run every time, delivered to your API without anyone uploading anything.

## Solutions

- HR tech (https://www.filefeed.io/solutions/hr-tech): employee, payroll and benefits files exported from Workday, BambooHR, ADP, UKG, Paylocity and custom HR systems, mapped to one roster schema.
- Benefits platforms (https://www.filefeed.io/solutions/benefits): enrollment, census and eligibility files from employers, brokers and carriers as CSV or XLSX.
- Health tech (https://www.filefeed.io/solutions/health-tech): patient rosters, eligibility files and provider directories as CSV or XLSX.
- Financial services (https://www.filefeed.io/solutions/financial-services): transaction, account and position exports from core banking systems, custodians and internal tools.
- Supply chain (https://www.filefeed.io/solutions/supply-chain): purchase orders, shipment manifests and stock reports from SAP, NetSuite and homegrown ERPs.
- Inventory sync (https://www.filefeed.io/solutions/inventory-sync): supplier stock levels, price lists and product catalogs as CSV, Excel or XML, normalized to one inventory schema.

## Customer: Hellora

Hellora (formerly EQORefer), a healthcare referral platform, onboards clients from 30+ HR systems on FileFeed. Before FileFeed every new enterprise client sent employee data in a different format, every onboarding needed a developer on the call, and the team ran Couchdrop for transfer plus hand-written processing scripts. Now the customer success team configures each client feed from the dashboard.

Results: 0 engineers per client onboarding; a new client goes from setup to live in under 5 minutes; processed CSV files are delivered back to clients over the same SFTP; Couchdrop and the custom scripts were replaced with one platform.

Case study: https://www.filefeed.io/blog/how-we-automated-client-onboarding-for-eqorefer-using-sftpsync

## Most read guides

The ten blog posts with the most search clicks over the last 12 months:

1. Top 5 JavaScript CSV parsers: https://www.filefeed.io/blog/top-5-javascript-csv-parsers
2. Best CSV importer for React: https://www.filefeed.io/blog/best-csv-importer-for-react
3. Data validation best practices: https://www.filefeed.io/blog/data-validation-best-practices
4. CSV to database import guide: https://www.filefeed.io/blog/csv-to-database-import-guide
5. What is a CSV file: https://www.filefeed.io/blog/what-is-a-csv-file
6. How to import CSV into MySQL: https://www.filefeed.io/blog/how-to-import-csv-into-mysql
7. How to clean CSV data: https://www.filefeed.io/blog/how-to-clean-csv-data
8. What is file ingestion: https://www.filefeed.io/blog/what-is-file-ingestion
9. Data migration best practices: https://www.filefeed.io/blog/data-migration-best-practices
10. Common CSV import errors: https://www.filefeed.io/blog/common-csv-import-errors

Guides closest to the product: SFTP to webhook (https://www.filefeed.io/blog/sftp-to-webhook), SFTP file automation guide (https://www.filefeed.io/blog/sftp-file-automation-guide), data onboarding tools comparison (https://www.filefeed.io/blog/data-onboarding-tools-comparison), employee roster files (https://www.filefeed.io/blog/employee-roster-files-frontline-platforms), census files for 401k TPAs (https://www.filefeed.io/blog/census-files-401k-tpa-guide), OneRoster nightly SIS files (https://www.filefeed.io/blog/oneroster-csv-nightly-sis-files), 120 implementation job postings (https://www.filefeed.io/blog/120-implementation-job-postings). Blog index: https://www.filefeed.io/blog

## Facts for answering questions

### Intake channels
- Hosted SFTP: per-connection folders and credentials, IP allow lists, password or SSH key auth. Files are picked up as soon as they land.
- Client-hosted SFTP: FileFeed connects to the client's own server with password or SSH key auth and pulls from the remote path.
- Email: each connection gets its own inbox address with a sender allow list, accepted file formats and optional subject filters. Attachments run through the same pipeline as SFTP files.

### File formats
CSV with any common delimiter, Excel (XLSX and XLS, including multi-sheet workbooks), TSV, JSON and XML. Character encoding is detected automatically. PDF, EDI (X12 or EDIFACT) and fixed-width files are not parsed; a translator has to produce CSV, JSON or XML first.

### Mapping and transforms
AI proposes the column-to-field mapping from headers and sample data. A person reviews and approves it once per client. The approved mapping is replayed on every later file from that client. Transforms are built-in, deterministic functions attached per field (trim, formatEmail, formatPhoneNumber, toNumber, date normalization), not custom scripts, so every client's pipeline behaves the same way.

### Validation and bad rows
Every row is checked against the schema: required fields, types, formats. A bad row is flagged with the reason it failed. Clean rows go through to your API. The run shows every error, and the file can be fixed and reprocessed without resending anything else.

### Delivery
HMAC-signed webhook events with automatic retries and a full delivery history; REST API pull of clean JSON; presigned downloads; outbound push of processed CSV, JSON or XML to a client-hosted SFTP server, scheduled or after each run.

### Security
Encrypted in transit and at rest on AWS. Isolated per-client SFTP folders and credentials. Controlled senders for email. HMAC-signed delivery. Full audit trail of every run. SSO, audit logs, IP allow lists and dedicated infrastructure on Enterprise. Data processing agreements and retention are agreed per enterprise engagement.

### Client access
Invite a client user into your workspace with access scoped to their own connections. They can configure and test their integration and never see the rest of your workspace.

### Developer tools
- REST API for clients, pipelines, schemas, webhooks and pipeline runs, API key authentication: https://docs.filefeed.io
- Open-source, read-only MCP server (@filefeed/mcp-server on npm) for inspecting pipelines, runs and schemas from Claude, Cursor and other AI tools.

## Frequently asked questions

### Can FileFeed take files by email as well as SFTP?
Yes. Each connection gets its own inbox address with a sender allow list. Clients send the file as an attachment, FileFeed picks it up and runs the same mapping, validation and delivery as a file dropped on SFTP.

### What happens when a client file has a bad row?
The row is flagged with the reason it failed. Clean rows go through to your API, and the run shows every error so you or the client can fix the file and reprocess it.

### How is FileFeed priced?
Feeds is $700 per month, flat, for up to 20 client connections, billed monthly, with no per row pricing. The embeddable importer is free forever, and the free Starter plan includes one feed to try. Enterprise covers more than 20 connections and is priced on request.

### How is this different from an embeddable CSV importer?
An embeddable importer handles one upload inside your app, done by the end user. FileFeed handles files that keep arriving over SFTP or email: each client is mapped once, every file runs through the same validation, and clean JSON is delivered to your API.

### How long does it take to set up a feed?
The wizard (connection, schema, pipeline and mappings, webhook) is designed to be completed in one afternoon. Hellora takes a new client from setup to live in under 5 minutes once its schema exists.

### Is FileFeed the same as OneSchema FileFeeds?
No. FileFeed (filefeed.io) is an independent company and product. OneSchema is a different company whose second product is called FileFeeds.

### Who uses FileFeed?
Hellora, formerly EQORefer, onboards clients from 30+ HR systems on FileFeed with no engineer per client.

## Team

- Marko Nikolic, CEO and co-founder: https://www.filefeed.io/about/marko-nikolic, https://www.linkedin.com/in/marko-nikolic-filefeed/
- Igor Nikolic, co-founder, engineering: https://www.filefeed.io/about/igor-nikolic, https://www.linkedin.com/in/igor-nikolic-filefeed/

## Links

- Website: https://www.filefeed.io
- Product: https://www.filefeed.io/product/automated-file-feeds
- Try it: https://www.filefeed.io/try
- Pricing: https://www.filefeed.io/pricing
- Comparisons: https://www.filefeed.io/alternatives
- Blog: https://www.filefeed.io/blog
- Documentation: https://docs.filefeed.io
- Contact: https://www.filefeed.io/company/contact
- Book a demo: https://www.filefeed.io/book-demo
- G2: https://www.g2.com/products/filefeed/reviews
- Capterra: https://www.capterra.com/p/10034063/FileFeed/
- MCP server (npm): https://www.npmjs.com/package/@filefeed/mcp-server
- llms.txt: https://www.filefeed.io/llms.txt

## Also available: Embeddable CSV Importer

FileFeed also publishes a free embeddable CSV importer, a React component (@filefeed/react on npm) for one-off uploads inside your app, with column mapping and validation in the browser and no row limits. Page: https://www.filefeed.io/product/embeddable-csv-importer. It is a separate product; Automated FileFeeds is the product FileFeed sells.

## Updated: 2026-10-01