Back to Who's it for

For data and ML teams

FileBackbone gives data and ML teams durable file history with branch workflows, so datasets and experiment inputs stay reproducible as teams iterate quickly.

Create account Read guide View HTTP API

What it is

FileBackbone is an API-first file platform that stores datasets and derived artifacts with commit history. Teams can branch for experiments, merge changes safely, and keep a clear lineage for reproducibility.

Track data and experiments with full history and context. Version datasets, reproduce results, and understand exactly what changed and why.

Why data and ML teams use it

Reproducible experiments

Keep exact dataset versions tied to outcomes so runs can be repeated with confidence.

Parallel iteration

Use branches for feature engineering and model prep work without stepping on each other.

Traceable lineage

Understand when data changed, what changed, and how it impacted downstream results.

Team-safe collaboration

Reduce accidental overwrites and keep shared datasets organized as the team grows.

Get started in 4 steps

  1. Create an account and a repository for your dataset project.
  2. Create an API key from the API keys screen.
  3. Initialize the client in your training or data prep pipeline.
  4. Write data updates with commit metadata and branch for experiments.
import { FileBackboneClient } from 'reposys/client/filebackbone_client.js';

const client = new FileBackboneClient({
  baseUrl: 'https://your-domain.example',
  token: process.env.FBB_API_TOKEN,
});

await client.create_branch(repo_id, { from_branch: 'main', new_branch: 'exp-feature-set' });
Follow the guide Browse JavaScript SDK methods

Keep data changes explainable

When your team moves fast, versioned history and branch-aware workflows keep results reproducible and decisions auditable.

Start building